The Impact of Generative Artificial Intelligence on Future

The Impact of Generative Artificial Intelligence on Future Economic Activities: A Strategic Trends Report

Yucong Duan

Benefactor: Shiming Gong

International Standardization Committee of Networked DIKWP for Artificial Intelligence Evaluation(DIKWP-SC)

World Artificial Consciousness CIC(WAC)

World Conference on Artificial Consciousness(WCAC)

(Email: duanyucong@hotmail.com)

1Introduction

Since OpenAI launched ChatGPT in November 2022, Generative AI has AI technology has achieved explosive development, attracting global attention and application boom. In less than a year, major technology companies have released powerful large models, such as GPT-4, Anthropic's Claude, and Google's PaLM 2, local large models such as Baidu's Wenxin Yiyan and Alibaba's Tongyi Qianwen have also appeared in the Chinese field. These models can automatically generate high-quality text, images, codes and other content based on natural language instructions, showing capabilities close to human creativity. With the improvement of technical capabilities, the application threshold of generative AI continues to decrease, and its influence is spreading from the technology circle to all levels of the economy and society.

It is worth noting that the speed of the global diffusion of generative AI is unprecedented. According to the World Bank's 2024 report, as of March 2024, the top 40 generative AI tools in the world have nearly 3 billion visits per month, with hundreds of millions of users, and ChatGPT alone accounts for 82.5% of the visits. Although this traffic is only one-eighth of Google's monthly visits, in less than 16 months, generative AI has spread to almost all countries in the world. Users are mainly young and highly educated groups, and tend to use these tools for activities that improve productivity, such as programming, writing, and design. Companies from all walks of life have also begun to try to integrate generative AI into their businesses: McKinsey's 2023 survey showed that 55% of organizations worldwide have adopted AI technology in at least one business department, a significant increase from the previous year; one-third of companies have already routinely used generative AI in marketing, product development, customer service and other functions. Corporate executives' understanding of generative AI has quickly shifted from ignorance to in-depth discussion in just half a year, and more than two-thirds of the companies surveyed plan to further increase AI investment in the next three years.

Looking ahead to the next 1 to 3 years, generative AI is expected to have a significant and far-reaching impact on human economic activities. This report adopts the perspective of strategic trend analysis and revolves around the following topics: First, how will generative AI change the organizational management and production processes of enterprises, including changes in process automation, document generation, project management, etc.; second, how will the labor structure be adjusted, which positions may be replaced by AI, which new occupations will emerge, and how will the demand for human skills be transformed; third, what changes will occur in consumer behavior, such as increased demand for personalized content, the popularity of AI-driven interactive customer service and shopping assistants, etc.; fourth, how will the industrial structure evolve, and whether generative AI will give birth to a new industrial ecology (such as AI content creation industry, AI agency platform, A I product chain, etc.); Fifth, from a macroeconomic perspective, what positive pull or negative impact will generative AI have on GDP growth, employment, inflation and economic resilience; Sixth, in terms of economic governance and ethics, issues such as data sovereignty, AI fairness, justice and the establishment of a regulatory framework need to be paid attention to; Seventh, combined with the network model of "Data-Information-Knowledge-Wisdom-Intention (DIKWP)" , we try to use semantic mathematics to remodel the information flow in economic activities, propose a generative economic model driven by DIKWP, and give examples of application scenarios such as AI-generated market research, consumer intention recognition, and enterprise knowledge decision-making systems. The report will finally summarize the main findings and look forward to the development trends in the next few years.

Please note that this report cites research data and cases from authoritative institutions such as McKinsey, the World Economic Forum, the Stanford University AI Index, and the World Bank to ensure the credibility and foresight of the content. In the text, we will present the key points in a concise manner, supplemented by detailed arguments and citations. Below, we will discuss them by chapter.

2.Changes in Enterprise Organization and Production Processes

Generative AI is driving profound changes in the organizational model and production and operation processes of enterprises. Its core is to act as an "AI assistant" or "co-creation partner" in various business scenarios to improve efficiency, reduce costs and stimulate innovation. Traditionally, many processes within enterprises rely on manual completion, such as writing documents, compiling reports, collating data, and formulating plans. These tasks are time-consuming, laborious and often trivial and monotonous. Now, with the help of generative AI tools such as Microsoft 365 Copilot, GitHub Copilot, Notion AI, etc., employees can hand over a considerable part of repetitive work to AI, so that they can focus on more creative and strategic matters.

Specifically, in terms of document and content generation, generative AI can automatically write business plans, meeting minutes, work reports, marketing copy, and other texts based on simple prompts, significantly reducing the time for manual drafting and editing. Some companies have taken the lead in deploying such tools and achieved remarkable results: for example, after an investment holding company introduced Microsoft 365 Copilot, the time for writing code was reduced from 8 hours to 2 hours, the cycle for developing chatbots was shortened from 3 months to 10 days, and the time for making presentations was shortened from 6 hours to 45 minutes. For example, after an Indian clothing company adopted the Copilot suite, it is estimated that employees saved 15% of the time they previously spent on administrative paperwork. It can be seen that generative AI significantly improves the output efficiency of knowledge-based work. In the field of software development, GitHub AI programming assistants such as Copilot can automatically generate code snippets based on natural language descriptions, helping engineers complete development and debugging tasks faster. It is reported that the use of these tools can increase development efficiency by an average of about 30%. IT companies such as Infosys said that the introduction of AI assistants has significantly accelerated the speed of feature development or vulnerability repair, and the quality of code has also improved. This means that AI not only speeds up the production process, but also reduces error rates and improves product quality.

In terms of process automation and operations management, generative AI can act as a digital assistant to optimize various business processes. For example, in project management, AI can automatically organize project information, develop Gantt charts, track progress, and generate minutes of regular meetings, saving project managers a lot of time on routine matters. An IT consulting company used Copilot to develop a "PM Assistant" tool, which resulted in a 30% increase in the speed of preparing project documents and a 60% increase in the efficiency of producing project launch briefings. At the same time, AI can also extract key points from project communication records to facilitate team knowledge sharing. In the customer support process, a generative AI-driven customer service chatbot can answer customers' common questions in real time, collect feedback, and intelligently upgrade complex problems to manual processing, thereby achieving 7×24 hours of online customer service and faster response. Many companies have reported that customer satisfaction and problem resolution rates have increased after introducing such AI customer service.

Collaboration and decision support are also important aspects of generative AI's transformation of corporate operations. By accessing the company's internal knowledge base, AI assistants can become employees' "intelligent search engines" and "think tanks." When employees encounter problems, they can directly ask AI to get answers or solution suggestions from internal documents and data, avoiding layers of reporting and lengthy communication processes. This helps break down information silos and enable knowledge to flow more efficiently within the organization. Some large multinational companies have developed internal GPT-type tools that allow employees to ask questions at any time to obtain the information and decision support they need while ensuring data security. For example, an American insurance group has developed a platform called "Secure GPT" that allows employees to use the powerful capabilities of generative AI to process text work while ensuring data privacy. Another example is that an Australian bank piloted Copilot to 300 employees and found that almost every user could discover multiple use cases to improve efficiency. The bank plans to further promote it in an all-round way.

In addition to improving efficiency, generative AI has also changed the innovation model and culture of organizations. With the creativity of AI, employees have a powerful assistant when brainstorming and designing products. For example, Alibaba integrated its large model "Tongyi Qianwen" into the enterprise collaboration platform DingTalk, where employees can let AI summarize chat records, generate meeting minutes, come up with company culture slogans, and even convert hand-drawn prototype sketches into mini-program prototypes. This greatly facilitates the precipitation of knowledge and the realization of creativity. AI can also provide new ideas for business decisions based on massive data and provide insights that humans may overlook. For example, Adobe integrates generative AI into its marketing cloud to help the market team automatically analyze consumer behavior data and generate insights, thereby formulating more accurate marketing strategies. It can be said that AI assistants are becoming "intelligence amplifiers" for organizational innovation, improving the scientificity and innovation of decision-making.

Of course, when companies introduce generative AI, they also need to make organizational changes and management optimization. In order to give full play to the role of AI, many companies have set up special AI empowerment teams to train employees and help them master prompt engineering. At the same time, companies are also adjusting processes to integrate AI outputs into existing business chains and establishing audit mechanisms to ensure the accuracy and compliance of AI-generated content. Overall, generative AI is driving companies to evolve in a more efficient, intelligent, and flexible direction: the degree of human-machine collaboration in daily operations continues to deepen, the trend of flattening the hierarchy and real-time decision-making is obvious, and corporate organizations are becoming more adaptable and creative. In the next few years, this transformation is expected to continue to accelerate in all industries, profoundly changing traditional work models and process designs.

3Adjustment of labor structure

As generative artificial intelligence penetrates into various fields, the human labor market is undergoing structural adjustments. On the one hand, some positions and skills are facing the risk of being partially replaced by AI; on the other hand, new occupations and skill requirements have emerged, giving rise to a new work model of "human-machine collaboration". Overall, in the next 1 to 3 years, the labor market will present a dynamic change pattern of "old positions transforming or disappearing, new positions emerging, and all employees upgrading their skills."

First of all, jobs that are easily replaced by AI are mainly concentrated in areas that are highly automatable and based on rules and document processing. For example, positions such as office clerk, administrative assistant, data entry, and customer service representative require processing large amounts of formatted information and repetitive tasks, which is exactly the strength of generative AI. Economists at Goldman Sachs predict that generative AI could replace up to a quarter of existing jobs, equivalent to about 300 million full-time jobs worldwide. Among them, office support and administrative clerical jobs are most likely to be replaced, followed by certain positions in the legal industry and some routine technical jobs in the fields of architecture and engineering. McKinsey's research also shows that by 2030, the demand for employment in occupations such as office support, customer service, and food and beverage services will decline significantly: for example, clerical and cashier positions may decrease by 1.6 million jobs, retail salespersons by 830,000, and administrative assistants by 710,000. The work content of these positions is programmable, and generative AI is capable of handling a large number of related tasks, such as document filling, form processing, customer Q&A, etc., thereby reducing the need for manpower.

However, it should be emphasized that "automation will replace tasks, not entire human occupations". Most occupations are not entirely composed of repetitive tasks, but include complex interpersonal interactions, creativity and strategies. Generative AI can often only replace the standardized parts, and humans will assume higher-value responsibilities in these occupations. For example, in the customer service industry, AI can handle common consultations, but human intervention is still required to handle difficult complaints, emotional comfort and establish customer relationships. In program development, AI can write basic code, but system architecture design, demand communication, final test optimization and other links still rely on experienced engineers. Therefore, many positions will experience "reshaping of responsibilities": the mechanical part of the work is handled by AI, and humans play more roles as supervisors, coordinators and decision makers. McKinsey proposed that generative AI will change the "anatomy of work" so that 60% to 70% of each employee's work content can be automated (higher than the previous estimate of 50%), but it is not common to completely replace the entire position. In fact, their model does not predict large-scale net unemployment, but predicts that by 2030, most occupations that are most affected by AI will still have a net increase in jobs, but the growth rate may slow down, and the content of the work will evolve significantly. This conclusion is also supported by historical laws: technological progress usually destroys some jobs, but also creates new jobs. In the long run, the emergence of new jobs often offsets or even exceeds the number of eliminated jobs.

So, what new jobs and skills are on the rise? With the popularity of generative AI, we have seen a number of new occupations enter the public eye. For example, "Prompt Engineer" "AI Engineer" has become a hot role. They specialize in how to design high-quality prompt words to get AI to give ideal results; "AI model trainer/fine-tuning engineer" is responsible for collecting and annotating data and adjusting model parameters to optimize AI performance; "AI ethicist" focuses on ethical, fair and just issues in the application of AI and formulates corresponding standards; there are also emerging positions such as "AI product manager", "dialogue designer" and "digital human operator". These positions did not exist a few years ago, but now they have rapidly grown into popular talents that companies are competing to recruit. According to statistics, in 2023 alone, 15,410 job recruitments were released in the United States requiring generative AI-related skills, of which 4,669 mentioned "large language models (LLM)" and 2,841 directly named ChatGPT experience. This number has increased by more than ten times compared with the previous year, highlighting the process of AI skills from scratch and the explosion of demand. Employers are increasingly favoring compound talents - talents who understand both business and AI principles and tools are sought after in the talent market.

Skill transformation and upgrading will become the norm in the wider workforce. The World Economic Forum's Future of Jobs Report 2023 points out that about a quarter of existing jobs around the world will undergo major changes in the next five years, and the demand for talents with AI and big data skills will continue to grow. The report predicts that the application of new technologies (especially artificial intelligence) and green transformation will create about 69 million new jobs, while about 83 million existing jobs may disappear due to economic pressure and automation. This means a net loss of about 14 million jobs, accounting for 2% of total employment, but at the same time, nearly 75% of jobs require employees to relearn new skills. Lifelong learning and on-the-job training have therefore become particularly important. Both companies and governments are aware of this and are taking measures to help the workforce adapt to the skills needs of the AI era. Some large companies have pledged to invest heavily in retraining their employees so that they can work with AI tools rather than be replaced; many online education platforms have also launched AI-related courses, covering a variety of topics from writing prompts, data analysis to AI ethics, for practitioners from different backgrounds to learn.

It is worth noting that the current trend is more towards "human-machine symbiosis" rather than "human-machine confrontation". A McKinsey survey shows that most companies tend to cope with the impact of AI through retraining and job adjustments rather than simply laying off employees. More than three-quarters of executives expect generative AI to bring about major industry changes, particularly in technology and financial services, with 75% of respondents believing that AI will significantly disrupt the status quo of their industry within the next three years. At the same time, only a minority of companies feel adequately prepared to manage these changes, with the most common concern being the accuracy of AI outputs (the “illusion” problem), but less than a third actually taking steps to mitigate the risks. This shows that on the one hand, companies realize that AI will have a profound impact on the labor structure, but on the other hand, more practice and exploration are needed on how to achieve a smooth transition.

Overall, the labor market will present a "double variation" in the next 1 to 3 years: low-skilled, automatable jobs may decrease at an accelerated rate, while high-skilled, creative and socially intelligent jobs will continue to grow. At the same time, each job will "evolve" due to AI - people need to master new skills to collaborate with AI and regard AI as an assistant tool to enhance their own capabilities. For individuals, continuous learning, improving digital literacy and adaptability are the keys to ensuring career resilience; for companies and society, formulating effective training plans and job transfer support policies, and cultivating a hybrid team of "AI+people" will help achieve inclusiveness in technological change and allow more workers to share the opportunities of the AI era.

4.Changes in consumer behavior

Generative AI not only changes the operation of enterprises on the supply side, but also profoundly affects consumer behavior and preferences on the demand side. As AI is increasingly integrated into daily life, consumers are showing new characteristics and trends in obtaining information, consuming content, purchasing goods, and interacting with enterprises. In the next 1 to 3 years, we expect consumers to pursue more personalized, interactive, and instantly satisfying experiences, and enterprises need to use AI to cater to and guide these changes and create a new consumption model.

First, the demand for personalized customization has increased significantly. For a long time, mass-market products and marketing methods that are all the same have been unable to fully meet diverse personal preferences. Generative AI has powerful data analysis and content generation capabilities, making large-scale personalization possible. Based on learning consumer behavior data and preference information, AI can adjust the products, services or content displayed to different users in real time. For example, e-commerce platforms use generative AI to dynamically generate product recommendations and advertising copy for each visitor, ensuring that marketing information fits the user's unique interests. Social media and streaming platforms use AI to create exclusive content streams or playlists for users, and the homepage that each person sees is customized and optimized. Consumers are becoming accustomed to this kind of “tailor-made” digital experience and tend to choose brands that can provide highly personalized services. According to the survey, when the content is in line with personal preferences, consumer engagement and satisfaction are significantly improved. Therefore, companies are turning to generative AI for precision marketing and differentiated product design. For example, companies such as Nike allow customers to design personalized shoes through AI; cosmetics brands use AI to generate customized product recommendations based on users' skin type and makeup preferences. From the perspective of consumer psychology, when consumers feel that products and services "understand them", their loyalty and willingness to buy will increase, which will in turn encourage more companies to invest in AI-driven customized experiences, thus forming a positive cycle.

Secondly, the way consumers interact with brands is more digital, intelligent and interactive. Traditional customer service often relies on human agents, with limited service time and high costs. Now, many consumers are accustomed to solving problems through intelligent customer service chatbots. Customer service systems supported by generative AI can simulate a conversational style close to that of real people, answer customers' questions in natural language, and handle after-sales requests. This type of AI customer service can be on call 24 hours a day, 7 days a week, respond quickly and not be irritable, greatly improving the customer experience. For example, OpenAI's ChatGPT has been integrated into website customer service by some companies to answer common product questions and guide operation steps. Another example is that the banking industry has launched AI financial advisors to provide customers with financial advice and product introductions through a chat interface. This interactive method is more personalized and accompanied, and sometimes customers can hardly tell whether they are talking to AI or real people. With the improvement of AI language model capabilities, this anthropomorphic interaction will become more natural and smooth. In shopping scenarios, AI shopping assistants are beginning to emerge: Amazon recently tested a generative AI shopping assistant named "Rufus" in its App. Users can describe their shopping needs to AI just like chatting with friends. AI will give product recommendations, feature comparisons and even purchase links based on Amazon's massive product information and online knowledge, and guide shopping decisions throughout the process. Imagine that in the near future, consumers will no longer need to browse and search hard when shopping online, but can directly ask AI "I want a pair of wireless headphones suitable for outdoor running, what do you recommend?" and then get professional advice and personalized options. This will greatly reduce consumers' information acquisition costs and improve the efficiency of purchasing decisions.

Immersive and interactive consumer experience is also a major trend. With generative AI, the boundaries between online and offline are further blurred, and consumers can enjoy a more realistic virtual experience. For example, in the fashion and furniture retail sectors, AI-supported virtual try-on/trial functions allow consumers to preview the effect of clothes on their bodies or the scene of furniture placed in the room in real time at home through their mobile phone cameras. This is equivalent to equipping each consumer with a portable designer and displayer, helping them to "see what you get" before purchasing, and enhancing their decision-making confidence. In the entertainment industry, the changes brought by AI are more colorful: streaming platforms use AI-generated content recommendations that are no longer limited to existing works, and may also generate customized plots or endings in real time based on the user's viewing history, providing an interactive film and television experience; in the game field, "adaptive games" with AI-generated levels and character dialogues have appeared. Every choice made by the player may be created by AI in real time. New plot branches, greatly enhancing the freshness and personalization of the game. There are even startups exploring the use of AI-generated "virtual idols" or "digital people" to interact with consumers, just like real-life Internet celebrities, live-streaming and selling goods, answering fans' questions, etc. These phenomena indicate that consumers are becoming increasingly accustomed to communicating with "digital avatars" and are willing to accept creative content and accompanying services provided by AI.

Content consumption behavior has also changed because of generative AI. On the one hand, the threshold for content production has been lowered, and a large number of AI-generated texts, pictures, and videos have emerged, and the amount of information that consumers can obtain has increased dramatically; on the other hand, consumers have also begun to pay more attention to the source and authenticity of content. For example, on news and social media, AI-generated texts and deepfake videos may be mixed in, and consumers have gradually developed a more cautious attitude and learned to distinguish the authenticity of information. This has prompted platforms to introduce AI detection and content tagging mechanisms to maintain consumer trust in content. In addition, the large amount of content created by AI provides consumers with more "long-tail" choices, and people's interest circles are more finely satisfied. However, there are also concerns that if everyone only receives personalized content selected by the algorithm, it may lead to information cocoons and aesthetic convergence. This requires content providers to strike a balance between personalization and diversity to ensure that consumers enjoy customization without losing the opportunity to explore new things.

In general, generative AI is catalyzing consumers’ move towards becoming “smart consumers”. The evolution of "generative consumers": they are more willing to use AI to obtain information and make decisions, and they also expect companies to provide real-time responses and products and services that are highly in line with their needs. The role of consumers has changed from passive acceptance to active participation and collaborative creation (for example, participating in customized product design or story plots). Faced with this change, companies must quickly adjust their marketing and service strategies and embrace AI technology to meet the expectations of the new generation of consumers. This includes deploying AI customer service, AI shopping guides, etc. to improve customer experience, as well as using AI to gain insight into consumers' potential needs and predict market trends, thereby leading new consumer trends. Companies that can make good use of generative AI to interact with consumers will win reputation and loyalty in future market competition; conversely, companies that cannot meet consumers' increasingly high digital expectations may be considered unattractive and gradually eliminated.

5Industrialstructure and emerging business formats

The rise of generative AI not only affects business operations and consumer behavior at the micro level, but also promotes the evolution of industrial structure at the macro level, giving birth to new formats and ecosystems. It can be foreseen that in the next 1 to 3 years, various industries will reorganize and innovate around generative AI: some industries will be given new momentum, some industries will merge their boundaries or even redefine them, and a new "AI industry chain" is taking shape, including basic model research and development, industry application development, platform service provision and other links. Let's analyze this structural change from several angles.

First, the AI content creation industry is emerging. In the past, the content industry (text, audio, video, design, etc.) mainly relied on manual creative production, which was costly and long. Generative AI can mass-produce content at low cost and high efficiency, making "content factory automation" possible. For example, marketing companies use generative AI to quickly generate social media copywriting, product descriptions, and advertising banners, greatly increasing the speed of content output. The film and television and game industries have also begun to try to use AI to generate character images, scene concept maps, and even first drafts of scripts. This has spawned a group of start-ups focusing on AI content production, such as Jasper (AI copywriting), Midjourney (AI image generation), RunwayML (AI video editing), etc. Some of these companies have received huge financing and become emerging content service providers. According to the Stanford University 2024 AI Index Report, private investment in generative AI-related start-ups soared to US$25.2 billion in 2023, accounting for about a quarter of all AI investments that year. Investors' enthusiasm for the "AI+content" track reflects the industry's optimism about the prospects of AI content creation. It is foreseeable that the traditional content industry will face a reshuffle: on the one hand, the threshold for content production will be lowered, and more individuals and small teams will use AI to participate in content creation, and the content supply in the market will grow explosively; on the other hand, human creators need to improve their skills and collaborate with AI to maintain differentiated competitiveness and provide creativity and depth that AI cannot replace. In the future, a new type of division of labor may be formed - a large amount of basic content is generated by AI, and human creators perform polishing and high-end creation, thereby achieving a balance between efficiency and originality. At the same time, the demand for derivative services such as content review, copyright confirmation and content curation (selecting high-quality content for users) will increase, becoming part of the AI content industry chain.

Secondly, AI applications and agent platforms have become new formats. Large general models (such as GPT-4 and Tongyi Qianwen) provide powerful general capabilities, but specific applications in various industries require customization and packaging of models. This has spawned the development of AI applications in many vertical fields, such as legal assistant AI (for reviewing contracts and finding regulations), medical assistant AI (diagnosis assistance, medical text summarization), and education AI (personalized tutoring, test paper correction). In order to facilitate enterprises and developers to deploy AI applications, AI platform services have emerged. For example, the APIs and cloud services provided by OpenAI, Baidu, and Alibaba Cloud enable others to build custom applications on their basic models. More and more startups are choosing the "Model-as-a-Service" model, opening interfaces for trained generative models through cloud platforms and charging by the number of calls. This platform economy has lowered the entrepreneurial threshold for AI applications and accelerated the penetration of AI in various industries. In addition, some AI agent platforms have also begun to emerge, allowing users to configure multiple AI agents to work together to complete complex tasks. Open source projects such as AutoGPT explore how to allow AI to autonomously call tools and disassemble tasks to form autonomous agents that collaborate with multiple agents. Although such platforms are still in their early stages, they herald a new business model: providing intermediary services for "AI labor." In the future, companies may no longer only hire human employees, but will also "hire" AI agents on demand from the platform to perform specific tasks as customer service, assistants, data analysts, and other roles. It is conceivable that a company may have a mixed team of humans and AI agents in the future, and flexibly expand or adjust the number and capabilities of AI agents through the platform. This will bring new challenges and opportunities to the traditional human resources service industry, and service providers specializing in managing the performance and reliability of AI agents will emerge.

Secondly, the "AI product chain" ecosystem is taking shape. Similar to previous technological revolutions, generative AI has also spawned a new industrial chain, which includes upstream computing power and algorithm providers, midstream model and platform providers, and downstream industry solution integrators. On the upstream side, training large models requires huge amounts of data and computing power, which has driven the rapid development of chip manufacturers and cloud computing services. GPU manufacturers such as NVIDIA have seen a surge in performance due to the surge in demand for AI training, and cloud service providers (such as Amazon AWS, Microsoft Azure, Google Cloud, Alibaba Cloud, etc.) are competing to launch high-performance computing clusters for large models for rent. It can be said that "shovel sellers" are the first to benefit from this wave of enthusiasm. On the midstream side, a group of "infrastructure" companies that provide general large models, as well as an open source community ecosystem, have emerged. For example, OpenAI, Anthropic, Baidu, iFlytek, etc. provide general large language models or multimodal models; Hugging Open source communities such as Face have become platforms for model release and communication. The Stanford AI Index report shows that 51 well-known machine learning models produced in 2023 are from the industry, and only 15 are from academia. The industry has become the leader in AI model innovation. The high cost of model research and development has also led to an increase in industry concentration to a certain extent. The funds and data scale required to train the top models are beyond the reach of small and medium-sized enterprises, and some underlying capabilities with strict technical barriers are mastered by a few giants. This has led to a situation of "basic model oligopoly and application innovation crowdsourcing": a few companies build common basic models, spread them to a large number of developers through APIs and open source methods, and they do a wide range of application innovations. This model is similar to the relationship between operating systems and app stores in the mobile Internet era, except that the "operating system" here has become a large model, and the "application" is a variety of derivative services around the model. In the downstream, leading companies in various traditional industries have combined with AI technology to form specific solutions for "AI+industry". For example, companies in the financial field have developed AI risk control and investment strategy generation systems; manufacturing has introduced AI for intelligent quality inspection and process optimization; pharmaceutical companies use generative AI for new drug molecular design and literature review to accelerate research and development. Different industries are also beginning to merge: for example, the automotive industry is combining with digital assistants to develop in-car AI assistants to enrich user experience; the media industry is cooperating with technology companies to launch AI news anchors, AI content review and other services. This cross-border integration has blurred industry boundaries to a certain extent, forming a new value network with AI technology as the link.

It is worth mentioning that the process of industrial restructuring is also accompanied by the evolution of fierce competition and cooperative relations. If traditional enterprises follow the wave of AI, actively invest in related technologies and adjust their business models, they are expected to improve productivity, reduce costs, and thus gain competitive advantages in the industry; on the contrary, those who stick to their old ways may be eliminated by the times. McKinsey estimates that the impact of generative AI on various industries will be different: industries such as banking, high-tech, and life sciences may gain new value equivalent to a few percentages of annual revenue due to AI. For example, the full application of AI in the banking industry can increase the value by 200 billion to 340 billion US dollars per year; the consumer retail sector can also increase the value by 400 billion to 660 billion US dollars per year. This indicates that there will be a gap between "AI leaders" and "AI laggards" within each industry. Those companies or regions that are the first to achieve AI transformation may expand rapidly, occupy a larger market share, and thus reshape the industry landscape. For example, companies in the retail industry have already used AI to achieve highly automated supply chains and precision marketing, thereby far exceeding their competitors in terms of profit margins. The regional pattern at the industrial level will also change as a result: countries with advantages in AI technology and talent will lead the upgrading of related industries, while regions lacking this foundation may further decline in the global value chain. Fortunately, both developed and developing countries are actively embracing AI. The World Bank report pointed out that the adoption rate of generative AI in middle-income economies is even higher than their economic size, contributing more than half of the global AI traffic, showing the potential of latecomers to "overtake on the curve" with the help of AI. However, this also depends on factors such as infrastructure and talent reserves. If countries can strengthen cooperation (for example, sharing best practices and developing interoperable standards), the emerging industrial ecology brought about by AI will be more prosperous and balanced. We have seen that some international organizations and alliances are committed to promoting an open AI innovation ecosystem, such as global collaboration in open source communities and joint release of AI application frameworks by multinational companies.

In short, generative AI is promoting a new industrial revolution. It breaks the boundaries of traditional industries and gives birth to new industries and new professions of "AI local" type; it also serves as a general enabling technology, deeply transforming the value chain of every existing industry. It can be foreseen that in the next few years, almost no industry will be able to be isolated from AI. In the face of this trend, various industries need to actively explore development paths that integrate with AI, while being vigilant to potential risks (such as excessive monopoly and unequal value distribution). Only in this way can we be invincible in this technology-driven structural transformation and jointly shape a healthy and diverse new AI industry ecosystem.

6Macroeconomic Impact

The large-scale application of generative artificial intelligence will have a multifaceted impact on the macroeconomy, including the increase in GDP and productivity, changes in employment and income distribution, the impact on inflation trends, and changes in the resilience of the economic system. In general, the "technological dividend" brought by AI is expected to become one of the important engines of future global economic growth, but the transformation process is also accompanied by short-term pain and structural challenges, which requires policy-level preparation.

First, in terms of GDP growth and productivity, various studies generally expect that generative AI will bring significant positive pulling effects. McKinsey’s latest research estimates that generative AI could add approximately $2.6 trillion to $4.4 trillion in value to the global economy each year, equivalent to the size of the UK’s annual GDP in 2021. If we consider embedding AI into existing software for more additional tasks, its economic contribution could even double. This increase is an additional 15% to 40% on top of the existing potential benefits of artificial intelligence, which means that generative AI is expected to become a key driving force for the next wave of productivity leaps. Specifically, AI can promote GDP growth through a variety of channels: First, improving labor productivity - AI automates some work, allowing workers to produce more per unit time. Goldman Sachs' analysis said that widespread adoption of AI is expected to increase the annual growth rate of US labor productivity by nearly 1.5 percentage points, which is a significant improvement against the backdrop of relatively sluggish productivity in developed economies. McKinsey's long-term forecast also shows that if AI technology is successfully diffused and the saved time is invested in other production activities, generative AI alone can contribute an additional 0.1 to 0.6 percentage points of labor productivity growth each year between 2020 and 2040, and the addition of all automation technologies can contribute an annual growth rate of 0.5 to 3.4 percentage points. The second is to give birth to new products and services - just as the Internet and smartphones brought many new industries in the past, generative AI will create new market demand. For example, emerging consumer areas such as personalized AI education, AI entertainment content, and smart home assistants will increase added value and boost GDP from the consumption side. The third is to improve total factor productivity - AI optimizes resource allocation and decision-making, making the combination of capital, labor and technology more efficient, thereby improving the overall output efficiency of the economy.

However, the macro effects are not evenly distributed over time and across different groups. In the short term and transition period, employment shocks and income distribution issues may become prominent. As discussed in the labor section above, AI may replace some jobs and reduce the demand for some low-skilled labor, thereby exerting downward pressure on the employment rate. If automation occurs too quickly and there are no supporting measures, some unemployed people will find it difficult to find new jobs in time, which may trigger a periodic risk of rising unemployment. Especially when the economy itself slows down or encounters a cyclical recession, the superposition of technological shocks and economic downturns will amplify job market fluctuations. The World Economic Forum report estimates that there will be a net loss of 14 million jobs in the next five years. Although this number is not large relative to the global labor market, the impact may be more concentrated in specific industries and regions. For example, workers in areas such as customer service centers and manufacturing assembly lines may face layoffs. This leads to the challenge of income distribution: the replaced jobs are mostly low- and medium-skilled workers, while high-skilled talents and capital owners who can master AI tools will receive higher income returns. If this trend is not intervened, it may widen the gap between the rich and the poor and the "digital divide." Historical experience shows that technological progress tends to increase overall income levels but exacerbates income inequality in the early stages, because the benefits flow more to people and companies that master new technologies. If the wealth brought by AI is highly concentrated in the hands of a few large technology companies and capital parties, while the income of a large number of workers declines or stagnates, then total demand (consumption) may be insufficient, weakening the sustainability of economic growth from a macro perspective. This is why many economists advocate that the government consider achieving a certain degree of rebalancing in the distribution of technological dividends through taxation, social security and other means to ensure that the fruits of growth benefit a wider range of people.

Regarding inflation, the impact mechanism of generative AI is more complicated. On the one hand, AI improves efficiency, reduces production costs, and has the effect of curbing inflation. For example, as companies adopt AI and their operating costs decrease, they can pass on some of the savings to consumers, slowing the price increases of goods and services. In the long run, AI-driven increased supply capacity (more goods and services being produced) will help ease the structural inflationary pressure caused by “supply and demand imbalance”. Automation can play a certain hedging role, especially in situations where labor shortages and rising wages drive up costs. In fact, in economies with a serious aging population (such as Japan and Western Europe), AI may become an important factor in maintaining the labor supply and avoiding a wage-price spiral caused by labor scarcity. On the other hand, AI may also bring new inflation drivers. In the short term, large-scale investment in AI (equipment, software, training) will increase demand and prices in certain areas. For example, the cost of high-end chips and cloud computing services will increase significantly. If technological changes lead to increased unemployment and reduced income, the government or central bank may adopt stimulus policies to protect employment, which will bring different price impacts. In addition, the popularization of AI makes some goods extremely cheap (such as digital content), but it may also give rise to new high-value goods (such as AI customized services) entering the consumer basket. These changes will be reflected in the weight adjustment of the inflation index. In general, most analyses tend to believe that the net effect of AI will tend to reduce inflationary pressure, similar to the past electrification and IT revolutions when productivity gains were often accompanied by lower inflation. But at the same time, it also reminds that policy authorities need to monitor price bubbles and asset market trends in AI-related industries to avoid the spillover of financial risks that affect price stability.

In terms of economic resilience, generative AI has both advantages and disadvantages. On the one hand, AI improves system efficiency and predictive capabilities, making the economy more responsive and resilient to shocks. For example, AI algorithms can monitor supply chain risks in real time, warn of potential disruptions (such as natural disasters and sudden outbreaks) in advance, and quickly adjust production plans to reduce losses caused by shocks; AI can also assist the government in making macroeconomic regulation decisions more timely, and through the analysis of massive economic data, identify emerging problems (such as economic overheating or recession signals) and make policy recommendations. Therefore, in theory, a highly AI-enabled economy should be more self-regulating and resilient than a traditional economy. In addition, AI can also help address some long-term structural problems and improve economic resilience. For example, in the context of an aging population, AI and robots can partially replace the labor force, so that the economy will not suffer a serious recession due to a shrinking labor force; AI can also optimize energy use and promote the development of clean technologies, so that economic growth is less dependent on fossil energy, thereby reducing the impact of sharp fluctuations in oil prices on the macroeconomy. On the other hand, we cannot ignore the new systemic risks brought by AI. When economic operations are highly dependent on several major AI systems or algorithms, if a failure occurs in one of them (such as an algorithmic error in decision-making, a cyber attack, or a serious "AI accident"), it may affect multiple industries in a short period of time, causing a chain reaction. In particular, as the financial market's reliance on AI trading and analysis increases, the probability of "flash crashes" and other phenomena may increase, and risk management needs to be strengthened. Similarly, AI may be used by criminals to create online rumors and financial fraud on a large scale, posing a threat to social stability and economic order. All of these require the construction of a risk governance mechanism in the AI era, including AI system redundancy design, network security protection, emergency plans, and legal liability definition, in order to enhance the resilience of the entire economic system in the face of AI-related risks.

Finally, from the perspective of the overall economic growth model, the widespread application of generative AI may push human society into a new growth paradigm. Some scholars have compared it to the paradigm shift brought about by "general purpose technology" (GPT, such as steam engines, electricity, and the Internet). Its characteristics include: growth is more driven by knowledge and intellectual capital rather than traditional material capital accumulation; the innovation cycle is shortened, the technology diffusion speed is faster but the competitive advantage is maintained for a shorter time; the importance of human capital is increased, especially the creativity, social emotional intelligence and other abilities that are difficult for AI to replicate have become the key to determining income. In this process, the relative positions of various countries' economies may also be readjusted. Economies that embrace AI early and successfully transform may seize the high ground of emerging industries and obtain higher potential growth rates; on the contrary, those who fail to transform may miss the technological dividend, fall into low growth or even be marginalized. International organizations and consulting agencies generally believe that if AI is properly deployed and supported by education and policies, the cumulative contribution of AI to global GDP will be considerable by 2030. For example, PwC has estimated that AI could contribute an additional $13.5 trillion to the global economy by 2030 (mainly from productivity and consumption effects). Goldman Sachs predicts that if the full potential of AI is realized, global GDP will be pushed up by about 7% in the next 10 years. Of course, the realization of this optimistic picture depends on many conditions, including investment scale, technological breakthroughs, social acceptance and international cooperation. If AI diffusion is not as expected due to poor supervision or too much social resistance, the macro gains will also be discounted.

Overall, the impact of generative AI on the macroeconomy is of great potential but full of uncertainty. The most likely scenario is that it will bring about a new round of technology-driven prosperity, improve productivity and create new demand, giving new impetus to global economic growth in the next few years. But at the same time, countries need to actively respond to challenges such as labor transformation, education upgrading and regulatory improvement to avoid economic fluctuations and social problems due to insufficient adaptation or lagging policies. In an ideal scenario, humans can control the pace and direction of AI development, making it both an engine of growth and a tool for inclusive sharing, thereby promoting higher quality and more resilient economic development.

7AI Economic Governance and Ethics

As the role of generative artificial intelligence in economic activities becomes increasingly prominent, the governance and ethical issues related to AI have become more important than ever before. Governments, businesses, and society need to strike a balance between encouraging innovation and preventing risks, and establish corresponding policies, regulations, and ethical norms to ensure the healthy development of the AI economy. In this chapter, we focus on several key issues: data sovereignty and privacy, algorithmic fairness and bias, and the construction of a regulatory framework, and discuss current progress and future directions.

7.1Data Sovereignty Sovereignty and Privacy Protection

Data sovereignty and privacy protection are the fundamental issues that AI governance must first address. The training of generative AI models often requires massive amounts of data, which may involve personal privacy, commercial secrets, and even national sensitive information. Therefore, governments around the world are generally concerned about how data is collected, stored, and used to maintain sovereignty and security. On the one hand, the EU’s General Data Protection Regulation The GDPR has made strict regulations on the use of personal data, requiring user consent, ensuring that data is used properly, and giving users the right to delete and correct data. ChatGPT was briefly banned in Italy for suspected privacy violations, which is a manifestation of the role of GDPR. OpenAI then took measures to add privacy protection options to resume service. On the other hand, China issued the "Interim Measures for the Administration of Generative Artificial Intelligence Services" in 2023, which is the first special regulation for generative AI and puts forward a series of compliance requirements for such services. The measures stipulate that companies providing generative AI services must conduct security assessments and algorithm filings in China, and the acquisition and use of training data must be legal and compliant, respect intellectual property rights, and must not violate regulations such as the Personal Information Protection Law. In particular, for AI services with "public opinion attributes or social mobilization capabilities" (such as chatbots that may influence public opinion), stricter security reviews and content supervision are required. These measures reflect the government's awareness of sovereignty over data and content: that is, AI service providers must comply with the laws and regulations of their jurisdictions and cannot allow cross-border data flows and algorithm black boxes to affect their own information security and cultural security. In addition, data sovereignty also includes the trend of data localization - some countries require that the data of their citizens must be stored on domestic servers and must not be transferred abroad without authorization. This poses a challenge to multinational AI companies, forcing them to adopt local deployment strategies in different markets or cooperate with local companies to obtain compliant data.

7.2Algorithmic Fairness and Bias

In terms of algorithmic fairness and bias, as AI’s involvement in decision-making increases, the fairness and non-discrimination of its decisions have become the focus of social attention. Due to the limitations of training data, generative AI may inherit or even amplify biases in the data. For example, some language models will unconsciously give biased answers when answering questions related to gender and race, which may affect user perceptions and even cause unfair results when applied to scenarios such as human resources and loan approval. Therefore, requiring algorithmic fairness has become one of the important principles of AI ethics. Regulators and industry standard organizations in various countries are working to establish fairness assessment criteria and accountability mechanisms for AI systems. For example, the European Union’s proposed “Artificial Intelligence Act (AI Act)” The AI Agency Law (AIAA) will classify some AI systems involving personal well-being (such as recruitment and credit scoring) as high-risk, requiring developers to prove that their models will not produce systemic discrimination, and to ensure fairness through transparency and human supervision. The Federal Trade Commission (FTC) has also repeatedly warned companies not to use biased algorithms, otherwise they may violate anti-discrimination laws. In the technical community, there are also many studies dedicated to bias detection and mitigation, including collecting diverse and fair data sets for training, calibrating model outputs, and introducing fairness constraint optimization. However, it is not easy to completely eliminate AI bias, because bias often reflects deep social and historical problems. Fair governance of AI requires the joint efforts of all parties, not only to improve the model technically, but also to increase transparency and accountability in management. For example, companies are required to disclose possible biases and limitations of AI systems to inform users; establish independent review agencies to conduct regular audits of important AI applications, similar to financial audits to ensure compliance and fairness.

7.3Construction of regulatory framework

The construction of a regulatory framework is currently the top priority of global AI governance. Governments are stepping up the formulation of laws and policies for AI to regulate corporate behavior and prevent risks without stifling innovation. As the White House science and technology adviser said, the current regulation of AI is a "race against the unknown" - technology is developing rapidly and regulation cannot lag behind. Overall, different countries have adopted diverse and different regulatory paths. The European Union is at the forefront, and its "Artificial Intelligence Act" is already in the legislative process and is expected to become the world's first comprehensive AI law. The bill adopts a risk-based framework, dividing AI systems into four levels according to their use and possible harm: unacceptable risk (such as social scoring systems, which will be disabled), high risk (such as recruitment, education, law enforcement and other key areas, which require mandatory compliance), limited risk (transparent identification of AI identity, etc.) and minimum risk (basically no regulation is required). For generative AI, the European Parliament proposed measures such as requiring such models to mark "AI generated" when generating content, and recording and disclosing the copyrighted data used for its training. This reflects the legislators' desire to ensure the traceability and transparency of AI output. In contrast, the United States currently has no unified federal AI law, relying mainly on the existing legal framework and guidelines of various agencies. For example, in October 2023, the US President signed the first executive order in the field of AI, requiring strict security testing of advanced AI models, promoting industry sharing of beacon watermarking technology to mark AI-generated content, and protecting consumer privacy. At the same time, more than 30 states in the United States have proposed their own AI bills, involving automatic decision-making tools, deep fakes, data protection and other aspects. It can be seen that the United States adopts a decentralized and flexible governance approach: the federal level gives principles (such as security, transparency, and non-discrimination), and states and federal agencies refine supervision in specific areas. The advantage of this model is that it encourages innovation and does not cut across the board, but the disadvantage is that it is prone to inconsistent standards and regulatory gaps. The United Kingdom chose to postpone legislation, and the existing industry regulators will act separately according to the AI regulatory principles issued by the government, and wait for the technology to mature before considering unified legislation. This "guidance first, legislation later" approach hopes to avoid the constraints brought about by premature legislation, but there is also a risk of insufficient enforcement.

In addition, China has adopted a strategy of active intervention in AI governance. In addition to the aforementioned generative AI management measures, China has also issued the "Algorithm Recommendation Management Regulations" and "Deep Synthesis Management Regulations" to form a relatively complete regulatory system. For example, deep synthesis regulations require that AI-generated media content (images, videos) must be clearly marked as AI-generated to prevent the public from being misled. In 2023, Chinese regulators continued to improve regulations and issued new draft rules for AI applications, which clearly require AI service providers to be responsible for the generated content and to promptly deal with any illegal and harmful content found. These measures show that the Chinese government hopes to keep a tight control on the "gate" in the early stages of AI commercialization to ensure that AI development does not deviate from the correct track (such as not endangering national security, disrupting social order, etc.), while promoting the credibility and controllability of AI technology through supervision. In contrast, many other countries are still at the research and consultation stage. At least 69 countries around the world have put forward more than 1,000 AI-related policy initiatives and legal frameworks. Despite different paths, countries have similar goals: to promote AI innovation while protecting security and privacy.

At the international level, the call for coordinated AI governance is also growing. Since AI itself and its impact are cross-border, it is difficult to cover all problems by the actions of a single country alone. In recent years, international organizations such as the G7, the United Nations, and the OECD have intervened in the AI governance agenda. In 2023, the Group of Seven (G7) launched the "Hiroshima AI Process" at the Hiroshima Summit, emphasizing the formulation of a responsible AI use framework. In November of the same year, the British government held the first Global AI Security Summit, convening countries and companies to discuss the governance principles of cutting-edge AI (such as superintelligence) and trying to reach an international consensus in the field of high-risk AI. As early as 2019, the Organization for Economic Cooperation and Development (OECD) issued the AI Principles (later adopted by the G20 and others), proposing that AI should be "people-oriented, respect human rights, transparent and explainable, robust and secure, and accountable" and other core concepts, which have become the basis of many national policies. In addition, UNESCO also adopted the "Recommendation on the Ethics of Artificial Intelligence" in 2021, calling on countries to incorporate ethical perspectives into the AI life cycle in legislation and practice. These international efforts show that global joint governance of AI is gradually getting started. However, as mentioned above, the definitions and regulatory models of different countries vary, which has led to companies facing fragmented compliance requirements in different jurisdictions. An important task of international coordination is to reduce this fragmentation, provide more clear and consistent guidance for multinational companies, and prevent "regulatory arbitrage" (i.e. companies take advantage of loopholes to move their businesses to less regulated places).

In addition to macro policies, corporate and industry self-discipline also play an important role in AI ethical governance. Technology companies have released AI ethical guidelines, such as Google's AI Principles (emphasizing social benefits, safety, avoiding bias, etc.) and Microsoft's Responsible AI Guidelines, and established internal AI Ethics Review Committees to review high-risk projects. In 2023, some leading AI companies (OpenAI, Microsoft, Google, Meta, etc.) publicly pledged under pressure to take measures to ensure AI safety, including red team testing of models, sharing AI detection tools, and reporting model capabilities to the government. Although these self-discipline commitments lack legal binding force, they can play a certain supplementary role before supervision is in place. Industry associations are also developing standards. For example, IEEE has released the "Guidelines for the Ethical Design of Automated and Intelligent Systems", and the ISO organization is promoting the standardization of AI management systems. These initiatives help to form soft law (soft law). law), that is, to establish common standards followed by the industry outside the law to provide flexibility for AI ethical governance.

AI governance also involves other ethical issues, such as intellectual property and copyright. Generative AI models often use a large number of works on the Internet as training materials, which has caused dissatisfaction and legal disputes among copyright owners. Artists and writers are worried that their works will be "learned" by AI and used for commercial generation without receiving any compensation. Some legal cases have already occurred, such as the Getty Images sued AI companies for unauthorized use of its photo training models. Governance in this area needs to find a balance between encouraging technological progress and protecting the rights of creators. Possible solutions include: establishing a copyright collective management mechanism, where AI developers pay fees to copyright organizations; or passing legislation to explicitly allow "fair use for algorithm training" and at the same time establish a compensation fund to help affected industries. Regardless of the method, resolving copyright issues is critical to the healthy interaction between the content industry and the AI industry.

In general, the economic governance of generative AI is a systematic project that requires a multi-pronged approach in terms of law, ethics, technology, and international cooperation. In terms of law, countries should improve AI-related legislation as soon as possible, clarify red lines and bottom lines, provide compliance guidance, and leave room for innovation. In terms of ethics, it is necessary to strengthen the sense of responsibility of AI developers and users to ensure that AI respects human values and serves human welfare. In terms of technology, promote the development of transparent and explainable AI technology and privacy computing technology to make AI more controllable and auditable. In terms of international cooperation, actively participate in global dialogue, seek convergence of rules, and avoid regulatory arbitrage and arms race-style competition. As the Secretary-General of the United Nations said: "We are at the starting line of AI governance, and action and collaboration must run ahead of risks." Only by establishing a sound AI economic governance framework can we enjoy the prosperity brought by generative AI while minimizing its possible negative impact and truly achieve "responsible AI innovation" .

8A new paradigm of generative economy based on the DIKWP model

After discussing the impact of generative AI on enterprises, industries, and the macroeconomy in the previous article, this chapter will try to explore a new economic operation paradigm driven by the DIKWP model from the perspective of information flow and knowledge creation. DIKWP stands for "Data-Information-Knowledge-Wisdom-Purpose", which is an extension of the classic DIKW (pyramid model), emphasizing the important role of goals and intentions in the process of transforming data into useful knowledge and wisdom. Simply put, the DIKWP model believes that when organizations make decisions, they not only need the accumulation of data, information, knowledge, and wisdom, but also a clear purpose to guide the direction. Based on this, we can conceive a "generative economy" framework in which each link from data to decision-making is highly empowered and connected by AI, thereby achieving an efficient closed loop of information flow in economic activities.

Let's first look at the meaning and traditional process of each element of DIKWP in economic activities:

Data: refers to all raw facts and values that can be recorded and stored. In economics, this includes market transaction data, consumer behavior data, production operation data, sensor data, etc. In the past, companies often relied on manual statistics or simple information system collection to obtain data, which was limited in volume and had poor timeliness.

Information: Organizing and processing data and giving it semantics turns it into information. For example, summarizing sales data into annual reports and categorizing customer feedback is the process of informatization. Traditionally, manual or basic software tools are needed to analyze data to obtain information carriers such as reports and charts.

Knowledge: The understanding and experience formed by further refining information is the regular knowledge that can guide action. For example, through years of market information analysis, the trend of changes in consumer preferences is discovered, which is knowledge. Knowledge is often stored in the brains of people or documents and is the intellectual asset of the enterprise.

Wisdom: Comprehensively utilize knowledge, experience and judgment to make wise decisions and plans. Wisdom is reflected in high-level decision-making and strategic planning, which is a comprehensive grasp of complex situations and choices made after weighing them.

Intention (Purpose): The motivation and goal behind the action is the starting point and evaluation criteria for driving the entire process. The intention of the enterprise can be profit, growth, customer satisfaction, etc., and the intention of the consumer can be to meet a certain need or achieve a certain value.

In traditional economic activities, the process from data to information, knowledge, and then to decision-making (wisdom) is often linear and human-dominated: enterprises collect relevant data based on strategic intent, analyze it into information, and a team of experts extracts knowledge. Finally, management makes decisions based on knowledge and their own wisdom. This process takes a lot of time and manpower, and there is friction (information transmission loss) and lag between each link. In addition, the purpose of the guidance of each link is not clearly quantified, and is more conveyed by experience.

With the introduction of generative artificial intelligence and semantic technology, all of this will change qualitatively. We can imagine a "networked DIKWP" system, in which data, information, knowledge, wisdom, and intention are no longer transmitted in a one-way manner, but dynamically interact and provide closed-loop feedback through AI. Specifically:

Data layer: Through the Internet of Things and digital platforms, massive amounts of data generated by economic activities are collected in real time and stored in the cloud. These data include not only structured data (such as sales and inventory), but also unstructured data (such as social media comments, customer service conversation records, and video surveillance images). Generative AI can understand and transform unstructured data, such as automatically transcribing customer comments and analyzing sentiment, thereby turning previously difficult-to-use "dark data" into a usable source of information.

Information layer: With the help of AI, data is cleaned, organized, and visualized to automatically generate various reports and insights. Market research that used to take analysts weeks to complete can now be automatically summarized into briefings by AI. For example, an AI system can scan industry news and social trends on the Internet every day, and combine it with the company's internal sales data to generate a market dynamics briefing for management reference. This is actually the prototype of AI automated market research. Similarly, supply chain AI can summarize warehouse and transportation data in real time to display inventory and logistics status information. Because AI has the ability to generate human language, this information can be directly presented in natural language or charts for decision makers to read and understand, eliminating the problem of information overload.

Knowledge layer: This layer emphasizes the formation and organization of knowledge. Traditionally, knowledge is often scattered in the minds of experts or documents. By introducing the knowledge graph With semantic technologies such as Graph and Ontology, AI can structure the information of enterprises and markets into a network-like knowledge graph. For example, a retail enterprise builds a knowledge graph containing nodes such as products, consumers, competing products, and marketing activities, and AI continuously updates the associations based on new data. This graph gives AI a certain "understanding" of the business. Based on this, generative AI is no longer blindly statistical, but can perform logical reasoning and common sense application. When a manager asks "Why did the sales of product A decline this month?", AI can find related possible factors in the knowledge graph (such as a large discount on competing product B, repeated regional epidemics leading to a decrease in store traffic, etc.), and give explanations and suggestions based on knowledge. This is far more valuable than a simple digital report because it provides insights at the knowledge level.

Intelligence layer: This is the decision support stage. AI with knowledge and reasoning ability can be regarded as the "think tank" of decision makers. It can not only answer questions, but also simulate scenarios, predict consequences, and assist humans in making decisions. For example, AI can simulate the impact of different marketing strategies on market share, or predict the price trend of raw materials based on historical knowledge, so as to recommend the best solution to executives. In some highly programmed decisions, AI can even execute automatically. For example, the pricing AI of e-commerce platforms adjusts product prices according to real-time supply and demand and competition information (knowledge), achieving similar "smart and intelligent decisions". Of course, humans still hold the final wisdom gate in major decisions, but AI provides unprecedented breadth and depth of information reference, reducing blind spots and biases in decision-making. It can be said that human-machine co-creation of wisdom becomes possible.

Intent layer (purpose layer): This is a new element in the DIKWP model and is also the key to ensuring that the system has a sense of direction in the generative economic model. Intent can be a company's strategic goal or a consumer's personal preference. By clarifying the intention, the AI system can optimize the output of each layer in a targeted manner. For example, a company with sustainable development as its intention will add environmental impact measurement to AI analysis, so that decisions will not only look at economic benefits but also carbon emission indicators. This is equivalent to introducing an evaluation function at the knowledge and wisdom level, allowing AI to optimize under multiple objectives. For example, if an AI shopping assistant for a certain consumer understands that the user's intention is to "save money as much as possible", then the information and suggestions it provides (such as recommending the most cost-effective products) will be different from the suggestions for another user pursuing the latest trends. Intent Graph (Purpose The concept of a graph can be introduced to represent the relationship between an organization’s goals, strategies, actions, and results. By making intent explicit, AI can be guided in the face of uncertainty—especially important because traditional data-driven models tend to lose reference in completely novel situations, and clear intent can serve as an anchor for AI decision-making.

In this meshed DIKWP framework, the layers are not one-way transmissions, but rather feedback and co-evolution. For example, intentions will affect the focus of data collection (if the goal is to improve customer satisfaction, the company will collect additional data such as satisfaction scores); the acquisition of knowledge will prompt decision makers to adjust their goals (new market knowledge may cause the company to change its strategic intentions); decision execution (the output of wisdom) will generate new data, which will flow back to the data layer, forming a feedback loop. Such a cycle can be partially or fully completed by AI, forming an adaptive generative economic model.

Here are a few specific application scenarios to illustrate how this model works:

AI automatically generates market research : Traditional market research takes weeks from defining the problem, collecting data, analyzing to reporting output. Now, AI agents can continuously crawl online sales data, competitive product trends, consumer reviews, and integrate key information into professional reports, and even directly generate conclusions that executives care about. For example, "Consumers' interest in green and environmentally friendly products has increased by 10% this quarter, and competitor X has launched a new environmentally friendly packaging that has caused some of our customers to churn out" - such insights can be automatically given by AI by combining various layers of information.

Identification and guidance of consumer intentions : In the retail and service industries, understanding the true intentions of consumers is the key to successful marketing. Through the DIKWP model, the company's AI system can convert consumers' browsing, evaluation, inquiry and other data into information and knowledge, thereby inferring their purchase intentions or potential needs. For example, a user searches for mountaineering equipment on an e-commerce platform many times and consults customer service about hiking routes. AI recognizes that his intention may be to prepare for an outdoor hike. So the system promptly pushes him an article on "Beginner Mountaineering Equipment List" (AI-generated content) and recommends a combination of related product discounts. This not only meets user needs, but may also inspire purchases (converting intentions into actions). In turn, users' clicks and purchases become new data feedback, verifying or correcting AI's understanding of their intentions.

Enterprise knowledge decision-making system : Many large enterprises are trying to build an internal "enterprise brain", which is to gather data and knowledge from the entire company, and use AI to analyze and support decision-making. This is the application of the DIKWP model: enterprises build their own large language models or knowledge graph systems to allow AI to master all aspects of enterprise operations. When management has to make a decision on a complex issue, they can ask the system questions and get a comprehensive analysis, such as: "With the current increase in raw material prices and slowing industry demand, how should we adjust the production plan for the next quarter to avoid inventory backlogs?" Based on real-time sales data (data layer), supply chain information (information layer), past operating experience (knowledge layer) and the company's profit goals (intention layer), the AI system quickly simulates the results of different scenarios and gives suggestions such as "It is recommended to cut production by 10%, focus on the production of the best-selling model C, and renegotiate price terms with suppliers . "

The generative economic model based on DIKWP has several notable features: real-time, customization, and self-learning. Real-time is reflected in the fast flow of data, information and knowledge, and decision feedback can be adjusted almost in real time (realizing the so-called "real-time economy"); customization means that the intentions of each decision-making entity (enterprise or consumer) are taken into consideration, and the output results are optimized for specific goals; self-learning is because AI continuously updates knowledge and model parameters through each decision result, making the system smarter with each use. From a macro perspective, if more and more economic activity units adopt this model, the entire economy will be more agile and efficient, and resource allocation will be more in line with dynamic needs, thereby improving overall welfare. However, it should be noted that a DIKWP-driven economy is not without risks: AI can make mistakes after all, and if the data is wrong (“garbage in”), it may lead to knowledge and decision-making biases (“garbage out”); over-reliance on AI decision-making may also cause humans to lose their intuitive judgment ability. Therefore, humans still play a vital supervisory and value-checking role in this closed loop.

Academics have begun to explore the formalization of the DIKWP model, such as constructing graph structure representations at all levels and studying reasoning methods under incomplete and inconsistent semantics. These studies lay the foundation for realizing a true "AI cognitive system". Some experts even suggest that this is one of the feasible paths towards artificial general intelligence (AGI): by giving AI the understanding and representation of data, knowledge and purpose, it can have cognitive and decision-making capabilities closer to humans. Regardless of whether AGI can be achieved in the short term, the DIKWP idea has important implications for the digital transformation of economic activities. It reminds us that data and algorithms alone are not enough. Clarifying goals, building knowledge and applying wisdom are the core of creating value. And generative AI provides exactly the capabilities that were previously lacking: it can process data efficiently, assist in knowledge production, and participate in decision-making to a certain extent. Therefore, by making good use of the DIKWP model or similar frameworks, we are expected to create an "intention-driven generative economy" - in this new paradigm, data resources are fully utilized, the link from information to decision-making is shortened, supply and demand matching is more accurate and agile, and economic operation presents unprecedented intelligent characteristics.

9Summary

In summary, generative artificial intelligence, as a breakthrough technology, will have a profound multi-dimensional impact on human economic activities in the next 1 to 3 years:

At the enterprise level, generative AI has reshaped organizational and production processes by acting as an intelligent assistant. It has helped automate a large number of repetitive tasks such as paperwork and programming, improved employee productivity, and promoted faster and better decision-making and innovation. The organizational structure of enterprises has shifted to a flatter and more flexible one, and human-machine collaboration has become the norm.

In the labor market, AI substitution and new job creation coexist. Some jobs that are mainly based on rule processing will be compressed or even replaced by AI, while new talents and jobs that know how to apply AI will rise. The overall employment structure will evolve from a pyramid shape to a "dumbbell shape" (high-skilled and service jobs will increase, and medium-skilled mechanical jobs will decrease). All people need to improve their digital skills to adapt to this change, and society must provide retraining and safety nets to smooth the transition.

In terms of consumer behavior, generative AI has led to higher expectations for personalization and interactive needs. Consumers are increasingly accustomed to customized recommendations and AI customer service, and their decisions are more dependent on the advice of smart assistants. Shopping, entertainment and other experiences are more immersive and interactive. Companies must use AI to accurately understand and respond to consumers' implicit intentions in order to win the future market.

In terms of industrial structure, AI technology has become a new engine that empowers various industries, bringing about industrial integration and reorganization. On the one hand, new formats such as AI content creation and AI platform services have emerged, forming a new industrial chain; on the other hand, traditional industries have been upgraded by combining with AI, which is expected to improve efficiency and create new value. Companies and regions that are the first to adopt AI will gain a competitive advantage, and the industrial structure will evolve in the direction of "AIization".

In the macro-economy, generative AI has the potential to drive a new round of productivity growth and become an important driving force for economic growth. But at the same time, the employment and distribution issues brought about by technological shocks need to be managed. If handled properly, AI will reduce costs, ease inflation, and enhance economic resilience in the long run; conversely, it may also exacerbate short-term pain or risks. The key lies in policy guidance to ensure that the benefits of technology are widely shared and do not trigger systemic risks.

In economic governance and ethics, the development of AI urgently requires the establishment of a sound regulatory and ethical framework. Data sovereignty and privacy need to be protected, algorithm fairness needs to be supervised, and AI decisions need to be transparent and accountable. Countries are actively exploring legislation and standards, and international cooperation is also starting, so as to ensure safety and fairness without hindering technological innovation. Enterprises should exercise self-discipline and implement AI ethical principles to jointly create a trustworthy AI application environment.

At the information and decision-making model level, we proposed the concept of a generative economic model driven by DIKWP, which connects data, information, knowledge, wisdom, and intention in the form of a semantic network. In this model, generative AI acts as an intelligent engine connecting various links, realizing closed-loop automation and optimization from data to decision-making. This provides a new paradigm of intelligent and purpose-oriented economic activities in the future.

It should be emphasized that the above impacts are not isolated from each other, but are interconnected. When enterprises apply AI to improve efficiency, it will affect the structure of labor demand; changes in consumer behavior will force the innovation of industrial models; and macro policies and governance will react on the intensity and methods of micro-subjects to adopt AI. Therefore, we must use systematic thinking to view the impact and opportunities of generative AI on the economy. This change is as profound as the industrial revolution and the information revolution, but at a faster pace. All stakeholders - governments, enterprises, workers, consumers, and researchers - should be prepared to actively participate in shaping the economic landscape of the AI era.

10Future Outlook

Looking ahead to the next 1-3 years, the economic transformation led by generative AI will enter a critical deepening period. Some phenomena that are currently only in the pilot and early stages will gradually become popular and have a wider impact. At the same time, new challenges and uncertainties will also emerge, requiring our continued attention and response. The following are some of the recent developments in this report:

More comprehensive industry penetration: In the next few years, generative AI will spread from currently concentrated areas (such as the Internet, finance, and marketing) to a wider range of traditional industries. Manufacturing may see AI-assisted design and process optimization; agriculture may use AI to analyze agricultural data to achieve refined planting and market forecasting; in the medical field, in addition to imaging diagnosis, AI will also make breakthroughs in drug development, personalized treatment plan formulation, and other aspects. This comprehensive penetration will drive productivity improvements in various industries to varying degrees. We expect that by around 2025, at least half of large and medium-sized companies will deploy generative AI solutions in multiple core business links.

New equilibrium in the labor market: After the initial adaptation period, the labor market will gradually find a new balance point for human-AI collaboration. Some jobs will disappear and be replaced by new ones, and the total employment level may remain stable or even rise slightly. However, the content and skill requirements of the work will be very different from today, and "AI literacy" will become one of the essential skills for almost all professions. The education and training system will accelerate reform, with colleges and vocational institutions offering more AI-related courses, and companies forming a normalized mechanism for improving employees' digital skills. The concept of lifelong learning will be deeply rooted in people's hearts, and personal career paths will become more diverse and flexible. A new phenomenon that may emerge is that "human-machine co-working teams" become standard configurations, that is, each team has AI agents participating in the collaboration, human members focus on creativity and judgment, and AI is responsible for data processing and daily affairs.

A qualitative leap in consumer experience: Faced with increasingly intelligent consumers, companies will compete to launch more personalized and immersive products and services. It is expected that by 2025, all major e-commerce platforms will have launched mature AI shopping assistants, significantly improving the convenience and satisfaction of online shopping. Some physical retail stores may also introduce AI shopping guides and smart fitting mirrors to integrate online and offline experiences. In terms of entertainment, interactive content based on generative AI may usher in a blowout - interactive novels, personalized music, AI in video games NPCs (non-player characters), etc., make every experience for consumers unique. The word-of-mouth economy will be further strengthened: AI will help consumers more easily find high-quality products and content that suit their tastes, while inferior or dishonest businesses will find it more difficult to escape the "eyes" of AI and the collective wisdom of consumers. In general, the era of consumer sovereignty will truly arrive because of AI, and companies must serve users with higher agility and customization levels.

Emerging ecosystems and business models: We expect that some new business ecosystems that are currently unimaginable will emerge. For example, an “AI creative market” may emerge, where individual developers can trade their own AI models or prompts in the market, and companies can purchase them on demand for specific scenarios, similar to App Store model, but the transaction object is AI capabilities. For another example, the "digital human economy" may rise - more and more virtual Internet celebrities, anchors, and customer service are driven by AI, and a special production and operation industry chain is formed behind it. Some startups may focus on the fields of "AI psychology" or "AI training" to help companies optimize human-computer interaction effects or customize AI personalities to meet the needs of different user groups. Traditional Internet giants will also adjust their strategies to integrate generative AI into their core businesses: search engines will be transformed into conversational answer platforms, office software will be fully embedded with AI assistants, and so on. These new models will have a lot of trial and error in the early stages, but the winners will eventually completely change the industry landscape.

Lagged improvement in productivity data: In the short term, macroeconomic data such as productivity improvement and accelerated GDP growth may not be obvious, because the full effect of technology usually lags. However, we expect that by around 2024-2025, the global productivity growth rate is expected to end years of stagnation and reach an inflection point. Especially in those industries and countries where AI is widely adopted, output efficiency will begin to exceed historical trend levels. This may partially offset some unfavorable factors in the macro economy (such as insufficient labor supply due to aging). In terms of GDP, if calculated globally by purchasing power parity, AI's contribution may increase global GDP growth by 0.2-0.3 percentage points each year, accumulating trillions of dollars in output in ten years. Inflation may be generally mild and controllable, but we need to be vigilant about price trends in local areas (such as the "AI commodity basket" outside of food and energy). The general tone is that signs of prosperity brought by AI will gradually emerge at the macro level, but the extent and timetable depend on the policy environment and adoption speed of each country.

Policies and regulations are gradually being implemented: The next 1 to 3 years will be a critical period for AI regulation to move from discussion to implementation. It is foreseeable that the EU’s AI The Act is expected to take effect in 2024-2025, and member states will begin to implement unified rules. In the United States, if the two parties can reach an agreement on certain AI issues, a framework federal bill may be introduced (for example, deep fakes, AI use regulations in key industries, etc.). China may further improve the existing interim measures, introduce formal regulations, and strengthen law enforcement inspections. At the international cooperation level, several multilateral initiatives or agreements may be formed, such as standards for AI system security testing, cross-border data sharing, and sovereignty consideration mechanisms. Enterprises will face clearer compliance requirements, such as the need to add watermarks to generated content, establish human review procedures for AI decisions, and regularly report algorithm impact assessments to regulators. In the short term, this may increase the burden on enterprises, but in the long run it will help reduce the social risks of AI applications and build public trust. As regulations are implemented, the AI industry will also enter a rational growth stage from the initial "surge", and compliance and technology will develop in parallel.

Social cognition and ethical consensus: The public's understanding of generative AI will be more in-depth in the next few years. From the initial novelty and excitement, it will gradually turn to a rational view of the pros and cons. AI ethics will become a public topic, and schools may begin to teach basic AI common sense and ethics (such as how to get along with AI, guard against information traps brought by AI, etc.). Public opinion will be more sensitive to the problems caused by AI. If there are safety accidents or mistakes caused by AI, it will promote calls for strengthened supervision. Correspondingly, successful AI application cases (such as AI helping new drug development to save lives, and AI education helping students in poor areas) will also increase public acceptance and support for AI. In general, human society will form a consensus on the role of AI through continuous dialogue and practice: regard it as a tool rather than a master, and emphasize the "people-oriented" AI use guidelines. We may see some form of "AI use guidelines" widely advocated, similar to the concept of green sustainability, gradually integrated into the social value system.

11Conclusion

The revolution brought about by generative AI has just begun. In the next 1 to 3 years, we will witness its gratifying results in benefiting the economy and improving welfare, but we will also inevitably experience some pain and tests. However, as long as we stick to the right direction - people-centered, prudent governance, open cooperation, and active innovation, there is reason to believe that generative AI will become a powerful boost to human economic development rather than a threat of getting out of control. Just as every technological revolution in history has ultimately promoted the progress of human society, this AI revolution is also expected to usher in a new golden age of productivity and creativity after overcoming challenges. Let us welcome the arrival of this wave of the times with a long-term vision and pragmatic actions, and create a beautiful future created by the collaboration of human wisdom and artificial intelligence.

人工意识与人类意识

人工意识日记

玩透DeepSeek：认知解构+技术解析+实践落地

人工意识概论：以DIKWP模型剖析智能差异，借“BUG”理论揭示意识局限

人工智能通识 2025新版 段玉聪 朱绵茂 编著 党建读物出版社

主动医学概论 初级版

世界人工意识大会主席 | 段玉聪

邮箱|duanyucong@hotmail.com

世界人工意识科学院

邮箱 | contact@waac.ac