Roundtable Dialogue: "Intelligent Leap", Unlocking the AIGC Innovation Ecosystem | WISE 2024 Business King
From November 28th to 29th, the two-day 36Kr WISE2024 Business King Conference was grandly held in Beijing. As an all-star event in the Chinese business field, the WISE Conference has now reached its twelfth edition, witnessing the resilience and potential of Chinese business in an era of constant change.
2024 is a year that is somewhat ambiguous and characterized by more changes than stability. Compared to the past decade, people are slowing down their pace and developing more rationally. 2024 is also a year in search of new economic drivers, and the new industrial changes place higher demands on the adaptability of each entity. This year's WISE Conference takes "Hard But Right Thing" as its theme. In 2024, what is the right thing has become a topic we are more eager to discuss.
As a cutting-edge field of the new generation of artificial intelligence technology, AIGC has ushered in a critical juncture for its development this year. Entrepreneurs are accelerating their layout, technology giants are actively entering the field, and local governments are also strongly supporting and promoting it through policy measures. From algorithm optimization to practical application, from financial support to ecological co-construction, AIGC shows great potential. As entrepreneurs persist in exploring the "difficult but correct things" and looking forward to the future of China's AIGC, a vibrant and infinitely possible AI innovation ecological landscape is gradually unfolding.
Against this backdrop, 36Kr, in collaboration with 36C, has specially planned a roundtable dialogue centered on "AI Innovation Map" and "AI Project Naming", with the theme of "Presenting the Largest Map of AI Innovation Ecology and Making Every AIGC Innovation Project Visible to the Entire Industry". The participants include Zhao Xi, Executive Chairperson of the Digital Economy Special Committee of the China Information Association, Cai Suqing, Founder of 36C, Grace, Strategic Advisor for Startup and Investment Ecosystem of Amazon Web Services, Geng Ruosi, Head of Startup Ecosystem of Volcano Engine Accelerator, and Zhang Yu, Co-founder of Aiebai. Five industry experts jointly explore AI innovation and its future trends.
Conference site
The following is the roundtable record, sorted out by 36Kr
Yu Lili: Good afternoon, everyone. I am Yu Lili, the main writer of 36Kr, and I have been focusing on the AI direction. The theme of today's roundtable is "The New AI Ecology UP Bureau". Currently, AI is a very hot track. Over the past year or more, 36Kr, in collaboration with 36C and several technology giants and large model companies, has been continuously promoting AIGC ecological cooperation.
Later, we will release an AIGC long list, which is designed and written by the 36Kr Research Institute. Currently, more than 100 companies have joined. The purpose of releasing this list is to showcase the AI innovation map and help more AI companies be visible to the entire industry.
Today, I am honored to invite guests from our partners, including Zhao Xi, Executive Chairperson of the Digital Economy Special Committee of the China Information Association, Cai Suqing, Founder of 36C, Grace, Strategic Advisor for Startup and Investment Ecosystem of Amazon Web Services, Geng Ruosi, Head of Startup Ecosystem of Volcano Engine Accelerator, and Zhang Yu, Co-founder of Aiebai.
Next, I would like to invite each guest to share their views on the AI innovation trend. Before speaking, you can briefly introduce yourself. First, let's welcome Chairperson Zhao.
Zhao Xi: Hello everyone. Originally, Chairperson Ni Chunyang was supposed to attend this meeting, but due to a temporary ministry meeting, he couldn't make it. As the executive chairperson, I am responsible for carrying out this meeting.
Let me briefly introduce the Digital Economy Professional Committee of the China Information Association. We were established in August 2018 and are the first national-level digital economy industry association in the country. It is supervised by the National Development and Reform Commission and operates under the guidance of the National Information Center. The reason for its early establishment is that from the end of 2017 to the beginning of 2018, the National Development and Reform Commission led the launch of a national-level strategy for the digital economy. Being within the system, we are well-informed and act quickly. When registering, we even registered the English name "Digital Economy Association" and the domain name "dea.org.cn" in advance, laying the foundation for our subsequent international exchanges.
In overseas exchanges, we represent China in promoting the development of the digital economy. For example, on December 7, we will organize a delegation of the "Belt and Road Digital Promotion Alliance" to conduct investigations in Singapore, Malaysia, the Philippines, and Sri Lanka.
There is also the latest information that on October 28, two ministries and commissions of the country issued a letter adjusting our supervisory unit from the original National Development and Reform Commission to the National Data Bureau. Currently, we are the only national-level industry association managed by the National Data Bureau, which is also very in line with today's theme.
As a national-level industry association in the field of digital economy, we are honored to participate in this work promoted by 36Kr and 36C. We believe this is a very meaningful attempt. Before participating, we carefully studied the ecological development laws of the industry. Usually, the development of the industry is driven by several mainstream manufacturers through technology to promote the market. However, we are pleased to see that this time, many mainstream manufacturers are jointly promoting one thing. This collaborative approach is very in line with our philosophy, so we actively participate in it.
Second, as an industry association that is very close to the government's will, our participation to some extent also represents the government's attitude. We hope to have closer interaction with the market. However, because the government's decisions often require a long period of brewing, this rigidity makes the government's actions relatively slow. Therefore, this participation is also an attitude indicating that the government is trying to change its own role positioning and try to make its participation more proactive and flexible.
At the same time, through the participation of our third-party industry association, since we are not manufacturers and do not represent any commercial interests, we hope that this platform and this list can be more open and fair. This is also some of my personal experiences in participating in this activity.
In the past year, regarding some changes in AIGC, I have some of my own understandings. Taking the call center and customer service scenarios that I am relatively familiar with as an example, in this field, the degree of human-machine coupling has exceeded 50%, that is, more than half of the work is completed by machines. Many years ago, everyone may still remember the scene of the 10086 call center. At that time, the business model was that a team leader led about 20 seat staff, usually young women. There would be a pre-work meeting before work, during work, the team leader was responsible for patrolling the field to check everyone's status, and there would also be a post-work meeting after work. This is a completely human-centered management scenario.
But now it is completely different. There are almost no manual seats in the 10086 front line, and the only remaining role is the artificial intelligence trainer. One trainer is responsible for about 25 machines, that is, more than 20 very intelligent machines. Therefore, the entire business scenario has undergone a huge change. In some basic scenarios, more than 50% of the business has been replaced by machines.
For this reason, I personally pay more attention to the roles of those people who cannot be replaced by machines in the human-machine coupling scenario. How should these people be trained? What is their role? We define them as "digital talents". Based on this, the Digital Economy Professional Committee of the China Information Association began to layout the cultivation of digital talents two years ago. Our first starting point is to promote the development of digital talents. These are some of our new explorations, and I would like to share them with you. Thank you.
Geng Ruosi: Hello everyone, I am Geng Ruosi, and I am currently in charge of the startup accelerator at Volcano Engine. Volcano Engine is the cloud service platform of ByteDance. On the one hand, we provide computing power and the ability of large models for the internal, and on the other hand, we also export these accumulated products and solutions to external customers to serve them.
From my perspective, a very gratifying change this year is that the overall invocation volume of large models has increased rapidly. It can be seen that more and more application scenarios are being discovered, and more and more people are beginning to think about how large models can truly bring improvements and adjustments to their businesses. This is a very obvious change.
Of course, there are also some unchanged things in this process. For example, one of the fastest-growing areas is the AI companion application. Many startups have emerged in this field. Although they are doing similar things, the markets, user groups, and age groups they target are very different. Some companies focus more on the domestic market, while others focus more on the overseas market.
From my perspective, even in the AI era, for products, the user insight behind them is still the key, and this has not changed. This may also echo today's theme: the truly "difficult but correct" things. Even with more and more AI opportunities, everyone still needs to focus on their core competitiveness and pay attention to the real business insight in the products. These are some of my observations. Thank you everyone.
Grace: Hello everyone, I am Minghan (Grace) from the Startup and Investment Ecosystem team of Amazon Web Services. I am very glad to have the opportunity to share with you today. Amazon Web Services is the pioneer and leader in global cloud computing, and we are also in a leading position globally in the AI wave. In China, we are deeply rooted in the local market, helping Chinese enterprises achieve empowerment in technology and AI, and at the same time assisting Chinese enterprises to go global and explore the global market.
Next, I would like to share some of my insights and views on AI, cloud computing, and future trends.
First, with the continuous iteration of computing power and algorithms, we believe this has laid the foundation for the wide application of AI. At the same time, AI is deeply integrating with industries, reshaping the business forms of thousands of industries. The combination of AI with big data, cloud computing, the Internet of Things, biotechnology and other fields has given birth to a number of excellent startups. In the directions of AI + healthcare, AI + education, AI + finance, etc., the traditional industries are being changed, which also creates huge innovation and entrepreneurship opportunities for the younger generation of entrepreneurs.
Second, open-source models will have wide applications, which greatly reduces the cost for enterprises to embrace AI and use large models. Taking Amazon itself and the open-source community Hugging Face that we support as examples, the user download volume and relevant research data show that more than 80% of model application manufacturers are using open-source models. This makes AI have the potential of being inclusive, allowing more small and medium-sized enterprises to participate in AI innovation, thereby promoting the explosive growth of AI applications.
At the same time, we believe that multimodal models will become the mainstream in the future. Compared to single-modal models, multimodal models will make applications more friendly, intelligent, and usable for C-end users. For example, apps like Doubao Large Model, Kimi, Baixiaoying, and Zhipu Qingyan, whether in image recognition, video generation, or voice interaction and conversation of AI Agent, are becoming more intelligent and the iteration speed is very fast.
Finally, the development speed of AI is very fast, and the resulting governance issues and the improvement of laws and regulations will also receive great attention. Whether at the government level or the enterprise level, how to deal with the ethical issues, privacy protection, and the rapid update of laws and regulations of AI will be the key to future development. This is also an important direction that enterprises and the government need to jointly promote. These are some of my views. Thank you everyone.
Cai Suqing: I am QQ, and that's what everyone usually calls me. I have been in the industrial incubation field for 18 years and can be considered a peer of Ruosi from ByteDance and Grace from AWS. I have only been an entrepreneur for more than a year, and what I am currently doing is also related to business incubation. Today's AIGC project long list is jointly released by my startup company and 36Kr. During the process of organizing the long list, we observed some interesting perspectives.
First, during the process of organizing the long list, we discovered 15 new AI innovation tracks and collected more than 2,000 new AI application and tool projects. From these samples, we summarized three characteristics.
The first characteristic is that the AI track is currently very crowded. Judging from the number of projects, a large number of new projects are emerging every day, but at the same time, there is a certain mortality rate. It is more interesting to look at it from the perspective of the founders: we found that many founders are quickly testing and correcting mistakes. If one project is not done well, they will quickly start a new project, so it is often seen that old faces appear with new projects.
The second characteristic is that the monetization model of the current AI innovation projects is not yet mature. The income of most projects relies more on contingency, relationship, or customized delivery, and there are relatively few projects with truly replicable and scalable benefits.
The third characteristic is that the age of this wave of AI entrepreneurs is significantly younger than that of the previous generation of Internet entrepreneurs, and at the same time, they are more pragmatic and simple. They know how to save money and use the least funds and AI tools to improve the company's efficiency. Moreover, unlike the previous generation of Internet entrepreneurs who pursued going public, this wave of entrepreneurs do not have an obsession with going public. They care more about whether they can build a small and beautiful company that is meaningful to society, rather than necessarily promoting the company to go public.
From my perspective, there is a very interesting opportunity here. On one side are the rapidly growing AI innovation companies, and on the other side are listed companies and large companies that need to conduct market value management through mergers and acquisitions. The current situation of the capital market, especially the secondary market, is rather difficult. Many listed companies and large companies are looking for AI companies with cutting-edge and growth potential. If AI is a certain opportunity, then mergers and acquisitions may become a very highly certain market opportunity. I hope my startup company can seize this opportunity in the future. Thank you everyone.
Yu Lili: The second question is, I would like to ask each of you to talk about from your respective positions what resources and support you can provide for these AI innovation companies? First, let's welcome Chairperson Zhao.
Zhao Xi: I mentioned a point earlier that AIGC has become more and more useful in the past year, and in some scenarios, it can completely replace manual labor, especially in the customer service industry that I have been engaged in for nearly 30 years, where this change is particularly evident. Just now when communicating with friends from Volcano, we also talked about a customer service application they launched based on a large model. I judge that the market positioning of this application may be between small B and large C.
For the large B market, the customer service system may require more complex exchange and callback mechanisms. For many entrepreneurial enterprises, especially the more pragmatic young entrepreneurs nowadays, they tend to use the latest tools and the most cost-effective methods to achieve interesting and efficient things. Such market demand is precisely the advantage of AIGC.
In the business scenarios I have observed, the current replacement rate of AI for manual labor is approximately 50% to 60%, and it may increase to 70% to 75% in one or two years. As for when it will reach 85%, or even higher, I am not sure. But I believe that when the replacement rate exceeds 90%, there will be a bottleneck. Because it is difficult for machines to achieve 100% replacement of manual labor, those talents who cannot be replaced will become more and more expensive.
Recalling the beginning, when the customer service industry was included in the strategic emerging industries, the gold content seemed not very high. Most of the practitioners in the industry were young people who received short-term training. However, with the development of AI technology, structured and framework-based tasks are gradually taken over by machines, and the requirements for the remaining talents are getting higher and higher. They not only need to understand human-to-human communication but also need to know how to collaborate with machines, which makes talent cultivation more complex.
In the past few years, we have focused on two directions: artificial intelligence trainers and all-media operations. In 2020, the Ministry of Human Resources and Social Security of the People's Republic of China approved 16 new occupations within one year, and two of them were personally promoted by me, which are artificial intelligence trainers and all-media operators. For this, we have compiled two textbooks, namely "Customer Service and Artificial Intelligence Trainers" and "All-Media Operations". Both of these books have been published in two editions, and currently, they have sold more than 60,000 copies and more than 80,000 copies respectively. Surprisingly, more than 80% of the readers are not from traditional industries but are undergoing career transitions. This further illustrates that training those talents.