After experiencing the cumulative integration and iteration of generative algorithms, pre-trained models, multimodal and other technologies, artificial intelligence has gradually broken through the field of traditional analytical artificial intelligence and ushered in the accelerated development of generative artificial intelligence. Less than a year after the launch of generative AI tools, its application has shown "explosive" growth, and it is expected that the technology will have a profound impact on business operations and economic and social development in the future, and is expected to start a new cycle of technological innovation and create trillions of dollars of value for the global economy.
History of artificial intelligence
With the successive launch of applications such as Dall-E2, ChatGPT, and StableDiffusion, generative artificial intelligence has become the focus of national attention. Just two months after its launch, the ChatGPT app amassed 100 million active users, making it the fastest-growing consumer app ever. Many generative AI applications have emerged since 2022, and 2022 is therefore known as the "first year of generative AI". It is expected that such models will have a profound impact on business operation models and economic and social development, and artificial intelligence is expected to open a new round of technological innovation cycle.
Looking back at the development history of the artificial intelligence industry, the evolution of generative artificial intelligence has roughly gone through three stages. The first is the use of machine learning for analysis and prediction in the first decade of the 21st century, and various machine learning techniques have developed rapidly, analyzing massive amounts of data through machine learning models, drawing conclusions or learning from the output information. During this period, machine learning was seen as a powerful AI tool, and companies used it extensively in data analysis, pattern discovery, and future prediction to automate tasks at a much faster and larger scale than ever before.
The second "decade" of the 21st century has entered the stage of visual and language processing through deep learning. The perception ability of artificial intelligence has been further enhanced, and deep learning has become a new research direction in the field of machine learning. During this period, the application of deep learning in computer vision helped search engines and autonomous vehicles better classify and detect objects. In speech recognition applications, AI voice assistants are able to interact with users in a more natural way.
Finally, with the cumulative integration of artificial intelligence technologies such as generative algorithms, pre-trained models, and multi-modality, the "big explosion" of generative artificial intelligence has gradually emerged. The GPT-4 language model developed by OpenAI marks a new stage in language-based AI applications. Generative artificial intelligence will usher in broad development prospects, and a number of "unicorn" companies have been born in the "track". The Boston Consulting Group predicts that the market size of generative artificial intelligence will reach at least $60 billion by 2025.
Application scenarios and their impact on employment in different business modules and industries
Generative AI scenarios
Generative AI application scenarios include text generation, image generation, audio and video generation, and "digital human" generation. Among them, the text generation application based on natural language processing is the earliest application in generative artificial intelligence, which can realize text content continuation, text style transfer, summary generation and whole text generation, and the related personalized text generation and real-time text interaction have broad prospects.
The technical scenarios of image generation are divided into image attribute editing, image local generation and change, and end-to-end image generation. Among them, the first two landing scenarios are image editing tools, which have been widely used and related products are relatively rich; End-to-end image generation corresponds to two landing scenarios: creative image generation and functional image generation.
Audio generation applications can be divided into speech synthesis and music creation. Speech synthesis includes text generation specific speech and speech cloning fields, and its Chinese to generate specific speech fields has a high degree of technical maturity, and speech cloning is of great significance to film, animation and other industries.
Video generation mainly corresponds to three major areas: video attribute editing, video automatic editing, and video part generation. Among them, video attribute editing is more widely used, which can greatly improve editing efficiency; Video auto-editing technology is still in the experimental stage; The principle of video part generation is essentially similar to image generation. It is expected that video generation is expected to become a medium- to high-potential scenario in the field of cross-modal generation in the future.
"Digital human" generation can be divided into "digital human" video generation and "digital human" real-time interaction. Among them, "digital human" video generation is currently one of the most widely used fields; The real-time interaction of "digital human" is mostly used in visual intelligent customer service, and more emphasis is placed on real-time interactive functions.
Impact on the operation of business modules
Generative AI has a positive impact on the operation of user operations, marketing and sales, software engineering, and product development business modules. According to McKinsey & Company, about 75% of the potential value that generative AI can provide is concentrated in four areas: user operations, marketing and sales, software engineering, and product development in 63 use cases across 16 business functions. Among them, the user-run business mainly uses generative artificial intelligence to improve user experience and increase customer service productivity. The application of this technology can not only increase the problem resolution rate per unit time, but also greatly reduce the time spent dealing with problems and the agent turnover rate. Crucially, generative AI improves the quality of service for inexperienced agents. McKinsey research estimates that applying generative artificial intelligence to user service business can improve productivity and save 30%~45% of current business costs.
Apply generative in marketing and sales operations
Artificial intelligence to improve personalized marketing, content creation, and sales efficiency. The technology creates personalized messages based on user preferences and behaviors, and produces content such as brand ads, headlines, product descriptions, and more. In addition, generative AI can be integrated into a variety of applications to provide higher-quality data insights, better target user segments, and determine appropriate marketing strategies. McKinsey research estimates that generative AI can increase the economic value of marketing productivity by 5%~15%. In addition to the direct impact on marketing productivity, generative artificial intelligence will also have a chain reaction, increasing sales productivity by 3%~5%.
In the business of software engineering, generative AI can be used as a coding assistant to speed up developers' work. The application of this technology can reduce some of the workload, such as generating initial code, code correction, refactoring, and generating new system designs, and can also improve the work experience of software engineers. A recent study found that developers who use Microsoft's GitHubCopilot software complete tasks 56 percent faster than those who don't.
The application of generative AI technology in product development operations can reduce development and design time and improve product simulation and testing processes. It is found that generative artificial intelligence can increase the product development rate by 10%~15% and shorten the product launch cycle.
The impact of generative AI on different industries and employment
Generative AI will have a significant impact across industries. According to McKinsey & Company, the retail and consumer goods, banking, pharmaceutical and medical industries were the most affected. Among them, in the retail and consumer goods industries, generative AI can automate key operations such as user service, marketing and sales, inventory and supply chain management, which can increase industry productivity by 1.2% to 2.0%, and is expected to create an additional $400 billion to 660 billion in economic value per year. The impact of generative AI on the banking industry is also enormous, with the use of generative AI technologies such as AI virtual experts, accelerated code generation, and mass generation of customized content increasing productivity by 2.8% to 4.7% and creating an additional $200 billion to $340 billion in economic value per year. Generative AI can dramatically improve the speed and quality of R&D in the pharmaceutical and healthcare industries, increasing overall industry productivity by 2.6 to 4.5 percent, and is expected to create an economic value of $60 billion to $110 billion annually.
Generative AI will also bring opportunities and challenges to employment in different positions. On the one hand, generative artificial intelligence will promote the intelligent upgrading of jobs, and some jobs will be replaced. According to the analysis of Goldman Sachs Research Institute, the intelligent automation ability of generative artificial intelligence has greatly improved work efficiency and reduced operating costs, and traditional positions in the United States and Europe will be affected by different degrees of artificial intelligence automation, and generative artificial intelligence can replace a quarter of jobs. Generative AI, on the other hand, will also create new occupations. "Ask Guest" allows people to use natural language as cue words to get information or create works by interacting with AI. In addition, a large number of new jobs will be generated in its related fields, such as artificial intelligence trainers.
The economic and social value of generative artificial intelligence
Generative AI could add trillions of dollars to the global economy. McKinsey & Company estimates that if the 63 generative artificial intelligence analyzes are applied to various industries, it will bring $2.6 trillion to $4.4 trillion in annual growth to the global economy, equivalent to the size of the UK's GDP. This prediction does not take into account all generative AI applications, and the economic impact of generative AI could double if applications that have not yet been studied are taken into account. Generative AI can greatly improve the labor productivity of the entire society, but only if the technology is coordinated with the entire production structure and working mode of the society.
Generative artificial intelligence combined with other technologies is expected to increase labor productivity by an average of 0.2%~3.3% per year between 2023 and 2040, of which generative artificial intelligence can increase labor productivity by 0.1%~0.6%, depending on the technology adoption rate and the deployment of employee time in different activities. In addition, employees need training to learn to master generative AI-related technologies, and some employees may change careers. If employee transformation and other risks are effectively controlled, generative AI can make a substantial contribution to economic growth and make the world more sustainable and inclusive.
Generative AI has become a field with a wide range of applications. In the future, with the continuous innovation of technology and the release of market demand, generative artificial intelligence will be more widely used in all walks of life, creating more value for the economy and society, and at the same time greatly changing the business operation mode and people's way of life. At the same time, the rapid development of technology will also bring new risks and challenges, such as intellectual property protection, security, technology ethics, environmental impact, etc. To ensure that generative AI technologies achieve high-quality, healthy and sustainable development, industry players, policymakers, and consumers need to work together.
*This article was published in Communication World
Issue 925 August 25, 2023 Issue 16
Original title: "The Impact of Generative Artificial Intelligence on Economic and Social Society"
End
Author: Data Research Center, China Academy of Information and Communications Technology
Zhang Yankun, Wang Xuemei, Wang Weiguo
Responsible editor/layout: Wang Yurong
Reviewed: Shu Wenqiong
Executive Producer: Liu Qicheng
