Roundtable Dialogue: AI is responsible for 99% of the sweat, and humans are responsible for 1% of the creativity Alpha Summit

Highlights:

1. The large model is not the direction of startups, and China's tradition is still the king of application, so the current startups should spend a lot of effort on application.

2. Startups also need to pay attention to several points: they must make some applications native to the intelligent era. It is necessary to be forward-looking in the design of technical routes. Maybe Day One has to think about the business model and how to make money, because every token costs money, so the economic model must keep up. It is necessary to integrate with China's macroeconomic trends and get out of Chinese characteristics. The team is also the most important.

3. The most important thing about AI is that it can not only make the digital human have a human-like appearance, but also give him a very natural ability to communicate, truly like a human. Digital humans can have diverse personalities, abilities, and roles.

4. The three main application scenarios of AI digital employees in the financial industry are the first is efficient content generation and improving operational efficiency, the second is intelligent virtual customer service and improving user experience, and the third is optimizing operation processes and innovating business scenarios.

5. The emergence of AI digital employees is not to replace people, but to make technology serve people, not the compression of human positions, but to realize the liberation of labor through efficient human-machine collaboration, so that human beings have more freedom and possibilities, and new jobs and innovations will also emerge.

6. Ethics has learned that there is no fixed answer today, how cruel human beings are, but they don't care what they should do, and now AI also has ethics, we only care about us setting ethical rules for AI, it can't hurt me.

7. Don't treat AI technology as a flood beast, when he really comes, we must learn how to cooperate with him, how to play the role of human originality, as long as it is accelerated enough, I think it has an equal effect on society.

On December 17, Chen Shi, investment partner of Frees Capital, Dong Zhisheng, chief architect of Xiaoice Solutions, and Professor Wu Guanguan, distinguished professor of East China Normal University, Changjiang Scholar and author of "From Metaverse to Quantum Reality", visited the "Alpha Summit" hosted by Wall Street News, and talked with Yang Wei, associate professor of the Department of Strategy and Entrepreneurship at CEIBS, to discuss the past and future of artificial intelligence.

Here's a look at the highlights of Wall Street:

1. What do you think about AI start-ups?

Yang Wei, Associate Professor of China Europe International Business School:

Good afternoon, and welcome to the Roundtable on the "Next" Industrial Revolution in Artificial Intelligence, which I am moderated by Wei Yang, Associate Professor of the Department of Strategy and Entrepreneurship at CEIBS. So here, we will not only discuss artificial intelligence technology, but also discuss the application of artificial intelligence technology, and what kind of impact such application will have on our economy and our society.

In the last 12 months, what we've seen is a burst of new generation AI technology. In the CEIBS classroom, we will say that the evolution of this technology is never linear, but indirect. But this time, we see that such a low-level technology, such a general-purpose technology, has been rapidly popularized and applied in a short period of time. The technology itself and our applications are developing very rapidly, which is really an unprecedented speed.

But at the same time, we also see that you have a lot of questions, where are our opportunities, what is the impact on our economy and society, including our laws, our ethical issues, how do we get along with such a smart machine, such an intelligent technology? These are all questions that you are very concerned about, and this is what I want to talk to the three of us today.

Introducing our three guests, we are fortunate to have Chen Shi, an investment partner of Frees Capital, who is at the forefront of investment. Dong Zhisheng, chief architect of Xiaoice Solutions, and Professor Wu Guanguan, distinguished professor of East China Normal University, Changjiang Scholar, and author of "From Metaverse to Quantum Reality", gave us an in-depth discussion on artificial intelligence.

So in fact, I still want to ask Chen Shi some questions about investment at the beginning, because after all, we are still a forum related to finance and investment this time, so in fact, we just saw that you have introduced a lot in your speech, and then we can see that artificial intelligence is really breaking out in all directions, from technology to application, then I believe that in fact, there are many start-ups that are also seeking opportunities in various fields, and there are also many companies that come to you to invest, so in this case, when you make investment decisions, which companies will invest and which companies will not investWhat was the most important factor in this decision? How do you look at these start-ups that are talking about AI?

Chen Shi, Investment Partner of Frees Capital:

Thank you. This round of generative AI is a big opportunity. In the United States, many people are doing AI entrepreneurship, and China is certainly thriving now. From the perspective of the direction of entrepreneurship, I think it is relatively clear that the current large model is definitely not the direction of ordinary entrepreneurs.

There is also a middle layer, which is the middle layer between the large model and the application, which seems to be squeezed by both ends of the large model and the application, and it is still difficult to run out of the current industry situation in China. Therefore, I think that the Chinese tradition is still the king of application, so we should spend more effort on application at present.

In fact, if you look at the development of China's digital economy, the Internet came from the United States, but China exported the best practices for application. China should be the country with the best digital economy and application development in the world, this is China's development paradigm, so I think application entrepreneurship is a good direction.

The first key point of application entrepreneurship is to seize the capabilities provided by the AI intelligent era and make some applications that are native to the intelligent era. As I mentioned in the previous topic sharing, it is necessary to fully focus on the new intelligent supply and artistic creativity supply to make a high-quality native new application experience.

The second point, which was also shared earlier, is to choose some non-consensus places to do difficult but correct things, which is a decision on positioning or business direction selection.

Of course, there are some other considerations, such as you need to keep a safe distance from the large model, and you have to have your own business depth, which is also very important. For example, if the large model is separated from the data, it will not work without digitalization, so we can find some scenarios that are not fully digitized, and where the large model cannot get the data, maybe it will be safer.

In addition, the design of technical routes can be more forward-looking. I've been telling our portfolio companies or startups that you have to be bold and use more advanced, immature technical architectures to make your applications, and you can conceive and design your applications according to the capabilities of GPT 5 and GPT 6, the so-called next-generation models, which is also a kind of foresight.

Secondly, I think that entrepreneurship in the era of large models, maybe day one has to consider the business model and how to make money, because every token this time costs your money, so your economic model must keep up.

Another point is that I think it is best to combine it with the trend of China's macroeconomy, such as the engineer dividend, or the general trend of going overseas, including a trend of biopharmaceuticals, the trend of consumption upgrading, etc., these are very Chinese characteristics, belong to the β of the times, and we must seize them.

Of course, the most important thing is the team, the technology must be good, the industry must also understand, the so-called "technology first, industry first", I think this is the most important, thank you.

Yang Wei, Associate Professor of the Department of Strategy and Entrepreneurship at CEIBS:

Ok thanks. Chen Shi just said some very useful principles, as soon as you say this, we will be able to know which enterprises, including where the opportunities are? Maybe I very much agree with what you said, for our Chinese enterprises, our genes are more in the application, and value creation is not only the technology itself, but also more value creation in the application side.

Do you have any interesting stories that you recently talked about at the end of your speech, do you have one or two cases of companies that you have invested in, and you would like to tell us about the story behind this, and what is the prospect of it?

Chen Shi, Investment Partner of Frees Capital:

Let me give you an example, last year's investment in special look technology. The positioning of Tekan Technology is to make new commercial video services on TikTok and Douyin in China.

There are a few points to note here, the first one, the team is two Ali's p9s who came out with a group of Ali's people. They were originally in charge of Taobao's official content platform, one is called Luban, and the other is called tapai, which are all Ali's official tool platforms for merchants. So they actually came out with this kind of cognition, and then entered this relatively emerging track, especially the track of overseas TikTok e-commerce.

It is a trend for merchants to go overseas here, and short video or live broadcast e-commerce is an innovative model in China, which may not be very good overseas, which is a new opportunity. There are fewer competitors in the new industry, and they are already very professional players.

Secondly, Tekan Technology uses AI technology to empower TikTok and Douyin merchants, which can greatly reduce their costs, anchors can use virtual digital humans, and commercial videos can be generated almost with one click. As I mentioned earlier, TikTok and Douyin have connected the creator economy with e-commerce sales, and as long as this tool can improve efficiency, whether it is production efficiency or conversion efficiency, it can produce commercial returns. Therefore, the track it chooses is relatively close to money, and it may have an advantage in terms of monetization.

And in fact, it is doing well now, and it has also achieved higher income in China, and it will start to do overseas markets recently. In addition to AI-enabled video production, it is also a closed loop with a data aggregation function. Therefore, the production of each content to the sales or interaction effect of the output can be attributed in time. As I mentioned earlier, in the case of Meituan, by making the open loop a closed loop, merchants can achieve refined operations.

Yang Wei, Associate Professor of China Europe International Business School:

You said two very important points, the first is that we need to have a relatively clear market, and the second is that we also need to have a semi-annual ecology with a relatively clear technology, so that his technology can be extended. So in the case you just mentioned, I think both of these points are very important.

Okay, then our last question to ask Chen Shi is our future development, as I just said, the development of this AI is really very fast, so with the rapid development of technology and the development of the market, in the next 1-2 years, what kind of changes do you think this trend in investment will be?

Chen Shi, Investment Partner of Frees Capital:

The biggest change in investment this year is that many head VCs are investing in large models, but as I just said, I think we are still following the large models, and the possibility of running out is large factories, or listed companies, etc., of course, the need for some national teams in the academy still exists, and there are still opportunities. Of course, I don't think a lot of them are startup opportunities. So then everybody is probably looking at the app, and I think the app is an opportunity.

In fact, it is difficult for me to say which industry is good or bad, the United States is a few industries such as law, finance, medical care and education, but the United States and China are not the same, so I don't dare to say which industries in China must be bad. In fact, the entire application layer, in fact, has not yet run out of the recognized so-called best practices at the application level at home and abroad, so we have to wait and see.

2. What are the characteristics and new trends of digital humans?

Yang Wei, Associate Professor of China Europe International Business School:

In fact, the process of our entrepreneurship is a process of continuous trial and error, continuous iteration, in which our venture capital and our VC power are still very important. Okay, thank you Chen Shi. Then I'm going to change the topic, we just talked about the application, now we will talk about a more specific application, then I have a series of questions for our Zhisheng.

Now we Xiaoice are doing some digital human work, in fact, this is a very interesting field. My first question I want to ask you is, what are the points of integration between this part of our digital human and the current wave of large models and large language models?

Because in fact, when we saw that when everyone was talking about the metaverse last year, digital humans were actually already a relatively popular track. We've seen a few of them, including virtual idols, that they're already using technology like Motion Capture from the previous generation. What are the new characteristics of this generation of technology, this generation of digital humans, and what are the new trends?

Dong Zhisheng, Chief Solution Architect of Xiaoice:

Okay, thanks to the moderator. Then on this issue, I will talk about some of our Xiaoice's understanding of this piece. Digital human is more concrete, making the machine more and more like a human, with human biological characteristics, human interaction ability, human creativity, and it can serve our scene and our life, which is a definition of digital human.

As for the combination of digital human and large model, our understanding is that the large model solves the connotation of digital human. It is based on a general large model, and then based on industry knowledge, based on AI agent, and based on its multimodal creativity, it builds the connotation of digital human.

What we Xiaoice did solved the appearance of this digital human, we through the fusion of multiple AI technologies, and then made this machine more human-like, while combining our real-time interaction engine, to achieve this real-time interaction ability of people and digital humans face to face.

And then the third point is that we can make digital humans have diversity through some configurations, or some personalized settings, with diverse personalities, diverse abilities, and diverse roles, and finally serve in the scene, which is roughly such a difference.

3. What is the application of digital humans in the financial industry?

Yang Wei, Associate Professor of China Europe International Business School:

Therefore, the most important thing for us in generative AI is to not only make the digital human have a human-like appearance, but also to give him a very natural ability to communicate, truly human-like. Yes, so many of our guests here today are from the financial industry, so in fact, everyone may be more interested, what kind of specific applications do we have in this industry and this scene of finance?

Dong Zhisheng, Chief Solution Architect of Xiaoice:

Well, we have recently served some leading customers in the domestic financial industry, such as China Merchants Department, China Merchants Bank, China Merchants Securities, as well as some large domestic state-owned commercial banks, such as Bank of China, Minsheng Bank, etc. In fact, in terms of the landing scene, there are mainly three major parts at present.

The first is to do efficient video content production based on digital humans. Because in the financial industry, you will be involved in the generation of a lot of videos. For example, such as investor education videos, such as some analysts or fund managers based on us, to do the interpretation of the market, the interpretation of policies, the interpretation of the market, the analysis of the industry and some other products, etc. Requires a lot of video production, and this is the first piece.

The second part is to combine digital humans, combined with our business processes, to provide this efficient customer service, and then we can combine our industry knowledge to create this kind of immersive and all-weather, multi-channel coverage of this efficient financial services. This is the second piece of content. The first two scenarios are some inputs from the customer-facing side of financial institutions.

In the future, we hope to create a 1:1 digital avatar for every colleague in a financial institution, which can replace us to go to meetings, do content production, do live roadshows, and do some other things. In this way, we can free up our own personal ability to do more valuable things. This is the first one.

The second type is to combine large models, AI agents, IPA, and our existing digital systems to create a personalized intelligent assistant in all fields of finance. And then in the future, we can imagine a scenario, in the future, the internal employees of our financial institutions, we face so many of our business systems, and then there will be an agent in the middle to help you, and then for example, we can communicate with our intelligent assistant in a word, and then the intelligent assistant says, okay, I understand, and then leave it to me. That's about it. These are the three main application scenarios we are now in the financial industry, the first is content production, the second is efficient intelligent customer service, and the third is an internal agent.

4. Will AI have a substitution effect on humans?

Yang Wei, Associate Professor of China Europe International Business School:

Thanks to the very imaginative these applications. While listening to you, I was thinking, maybe in the future, my bank account manager, I want to contact him to check the content of my own account, that is not necessarily this real person serving me, it may be this digital doppelganger serving me, but he may be more efficient, I can find him 24 hours a day, can meet one of my needs.

I still find this very interesting, very imaginative, and very promising application in the financial industry. The two points you just mentioned are actually very interesting, that is, the front is that we do propaganda externally, our digital people, we do videos, and the internal ones are actually our digital people, which are actually replacing the functions and work of some existing employees.

So this question is going to be given to you, how do you think we will get along with AI in the future and how will we coexist in the organization? So will the virtual digital humans of AI really replace the employees of our banks and then serve customers, will there be such a substitution effect?

Dong Zhisheng, Chief Solution Architect of Xiaoice:

In fact, this topic is not only in the financial industry, but also in recent years, with the increasing ability of AI, everyone is discussing the question of whether AI will replace people, or even eliminate people.

In fact, in our understanding, technology is an empowerment for us human beings. To put it more in detail, AI can replace human employees to do some repetitive, inductive, and quantifiable work, and then liberate our energy.

I participated in a forum before, and one of the guests at that time said something that I thought was particularly meaningful, he said that the future is the model of efficient human-machine collaboration, where AI will help us bear 99% of the sweat, and we will be responsible for the 1% of the creativity, which is probably such a collaborative work.

For example, for example, some of the top analysts and fund managers in the financial industry as we talked about earlier are actually very valuable in their time, and they need a lot of time to do industry research, do some research reports, and so on. But now in order to promote his products or share his investment ideas with the majority of investors, he needs to spend a lot of time doing live broadcasts and roadshows.

But if there is a digital clone to take on such a role instead of him, let him focus on thinking, focus on industry analysis, so that the two are added, one is to do what he is more good at, and the other is to hand over the work that can be liberated to his own digital clone, so that 1+1 is greater than 2.

5. How to understand the ethics of AI?

Yang Wei, Associate Professor of China Europe International Business School:

Now everyone is asking, will AI replace humans? I very much agree with you, in fact, I think this is the release of our productivity and creativity, we don't have to do these repetitive tasks, but focus our energy and wisdom on how we do creativity, how to expand, how to explore.

I think this is a very promising estimate and expectation for the future, which is very interesting. In fact, we have now begun to move from the application of technology, to the opportunity to invest, to an impact on our entire society and economy.

So my next wave of questions will be given to our Professor Wu to discuss, we may have gone beyond the scope of our business, more about the impact of this artificial intelligence on society, which everyone is actually very concerned about, then in fact, the first question I gave Professor Wu is our regulatory problem, that is, in fact, since the outbreak of this wave of ChatGPT, what everyone has been asking is how our artificial intelligence is regulated, including our data privacy issues, including some of our artificial intelligence-generated content, which has a different information effect。

So how do we control this artificial process, or manage the distribution of AI-generated content? It's really a regulatory issue. On December 9th, the European Union launched the first law on the regulation of artificial intelligence, so how do you think that our ethical framework, or the rules of our society, its thinking, its logic, should be formulated in this regulation?

Prof. Wu Guanguan:

Thank you very much, Mr. Yang. First of all, I would like to respond to Mr. Dong's wonderful point of view. According to Mr. Dong's statement, our 1+1 is greater than 2, and the repetitive work is handed over to the agent, and we do creative work. I couldn't agree more, I think it's great.

But if we look at today's human society, the vast majority of people are doing some repetitive and not very intellectually involved work, what should they do? How can we make these friends still have a place in the future society? I am doing education and training in the university, and I know that this is very difficult, and the university is still divided into 985, and it is divided into one or two books, and many people can't turn to the creative track, so what should I do? So the situation is still quite complicated.

Returning to the question raised by Mr. Yang, I think that today's social reality is very complex, and the issue of ethical framework is also very complex. I'm looking at Twitter every day, and we've seen the whole EU break heads, including the current factions, one called the "Doomsday" (artificial intelligence is coming, we must control him). The other one is called the "Accelerator". Like Mr. Chen, I think it's a bit like an accelerator, he's not so worried. Are we only having these two factions today?

I think the problem is more complex, and there is a third analytical framework. For example, what is the ethical framework? What is ethics? Ethics never says what you should do, ethics always points to yourself, how I should act, this is called ethics. Confucius said, "Do not do unto others as you would have them do unto you," and this is about how I should act. I don't like it, how can I do it for Mr. Dong?

Kant said that I should act in such a way that the law on which I am based may become a universal law. Therefore, the ethics given by Kant is very strict, he said that you can't lie, because you lie, I lie, Mr. Dong also lies, and we Mr. Chen also lie, and this lie is meaningless. If you lie, others still don't believe you, and the lie itself cannot be generalized.

There is still no fixed answer to ethics today, and there are still many scholars who do ethics today. We've been talking about what I'm going to do for thousands of years, and I feel like all of a sudden we're all lying flat. We don't care about how we act, we only care about how we set ethical rules for AI, how it should act, and it can't hurt me.

For example, two years ago, we knew that there was still "I can't breathe" and "Black Lives Matter" in the United States, and we still had people on the streets, and you shot me with a gun. Right now, Hayi is doing to each other in the most cruel and undesirable way. Human beings are so cruel to each other, who cares about ethics? Does the European Union care? Human beings are in a mess, but human beings don't care, anyway, we can be cruel to each other, we can not be ethical, but I'm sorry, it seems that there is AI on the horizon now, and we ethicists have found that my work suddenly has a second spring.

Thousands of years can't be solved, and I can't solve it, but let's "ethically" set that AI should be good to me, that AI can't do this, that can't do that, and that only we can continue to be cruel to us. So I think there's a change in our time, and humanity has been rotten.

During World War II, friends who watched "Oppenheimer" knew that if a scientist made something for you, you really threw it away from a politician. It's cruel. But we are still arguing that this cannot be done, we humans cannot be so cruel to each other, but today all kinds of cruelty have been placed in front of our eyes, there is only AI in the eyes of ethicists, there is only AI in the eyes of EU politicians, it seems that only AI is cruel to us, so people are starting to rot.

That's why I've been saying that today we're looking at the matter of "regulation" in reverse: human beings are going to regulate, is it good to regulate? What does the code of ethics mean? If we don't even know what ethics is, we just know that others have to talk about ethics. Today it's the turn of AI, we say AI should talk about ethics, what about us?

So today I think the problem is very complicated, and the whole reality of our society has to be looked at in the same way, and it is becoming more and more that we call it impossible. We sometimes call it "singularity approaching". What does the singularity mean? One is the technological singularity, and the other is that our human civilization has reached the singularity itself, and I think we need to do a comprehensive reflection today.

6. Will AI accelerate social injustice?

Yang Wei, Associate Professor of China Europe International Business School:

Wow, this answer is really different from what I thought, and we really answered the problem of our supervision from Professor Wu at a very deep level. Therefore, artificial intelligence is not only about the changes in our economy, but also about our human beings' thinking about ourselves and our world, which is a very interesting phenomenon, and a very profound phenomenon, thank you very much Professor Wu for this answer.

Then your answer actually leads me to another question. In fact, we have just talked about it, including Chen Shi just said, let's just say that our technology, especially the technology of the large model itself, is beneficial to the giants, and the last to run out must be the giants.

So the other question that people are more concerned about now is whether AI technology will accelerate the injustice of our society? Will it lead to a greater tilt in the technology and capabilities of our society as a whole, including the distribution of our resources, including our entire distribution of wealth, rather than a balanced and equal process. So what do you think about that? Is it going to accelerate the process of such a social injustice?

Prof. Wu Guanguan:

This one is actually quite complicated. From an empirical point of view, we actually have every time that as long as the speed of technological development is fast enough, it can bring a more equal technological equality to society.

As long as this time is fast enough, for example, the first batch of enjoyment, for example, there is a medicine that allows us to live to be 120 years old now, I think Mr. Chen may have taken it a little earlier than us, he is an investor, and this may be the research and development of the pharmaceutical company he directly invested. But as long as the technology is fast enough, Mr. Dong will eat it immediately. I also guess I'll be able to eat it right away.

So in this sense, the most fearful thing is that we stop moving, and then all of a sudden we find the privileged class, and find that the good things are just that, there will be no more, I just need to grasp this thing, just like the previous territory, I just need to occupy this mountain, I can occupy this mountain for thousands of years, I can pass it on from generation to generation, this mountain is my open, this means the stagnation of technology.

So I personally think that there is really no need to treat AI technology as a flood beast, when it really comes, we have to learn how to cooperate with it, how to play the role of human originality, as long as it is accelerated enough, I think it has an equal effect on society.

There is a great sentence in Mr. Chen's speech just now, and I have memorized it, calling for learning Xi way they learn Xi. In fact, there are some things that we can't learn Xi, it can directly share weights, share gradients, and directly share knowledge in a very instantaneous, we can't, we learn from each other in this simulated way Xi we may only learn 70%, and it is discounted.

But by understanding how they Xi, we can redefine where our intelligence is, how we function, for example, how we can play an original role, and so on. So in that sense, we're still going to have a chance. So I personally don't think too much about technology as a flood beast, as long as it is accelerated enough, I think it has an affirmative effect on society.

Yang Wei, Associate Professor of China Europe International Business School:

So we have to be optimistic about technology, we have to be optimistic about the interaction between people and technology, which is not only the equality of people in our social center, but also the equality of people and technology, and we also have a lot of space to Xi learn from technology, so we have talked about how we should look at our overall technology and the change of our overall development from the beginning, from the application of our AI technology. Therefore, I would like to thank the three of us for bringing us very deep thoughts and discussions today. That's all for today's roundtable discussion, and thank you very much.

This article is from Wall Street News, welcome to download the APP to see more