Dialogue with Yu Guoming: Liberal Arts Students, How to Find Value in the AIGC Era

Source: "True Story Lab" (ID:zhengulab), author: Lin Qiuyi, editor: Gong Zheng

Image source: Generated by Unbounded AI

In the era of AIGC, a large number of liberal arts posts such as writers, painters, and planners face the risk of being replaced by AI. At the same time, the "intelligent emergence" of generative AI (meaning that AI can learn what humans have not taught AI) has also made people anxious about how to manage AI.

"True Story Lab" talked with Professor Yu Guoming, Director of the Academic Committee of the School of Journalism and Communication of Beijing Normal University, and asked him to share the future of liberal arts students (especially news or content people) and the new way of life of future human beings under the wave of AI.

Yu Guoming's core views are as follows:

1. In the era of AIGC, understanding technology and society, understanding the market and understanding design principles is an important direction for future new students and even liberal arts students to enhance their social value and market value. **Students of liberal arts and science are the "two wings of a bird" of social development, and the key to the future is to realize "human-machine collaboration and integration". **

2. **The most important function of generative AI is "human enhancement", which greatly reduces the difference in ability between the so-called elite and ordinary laymen. **This kind of human enhancement eliminates the gap caused by human innate talents, and will construct the interactive relationship and corresponding model of the future society.

3. In the era of AIGC, it is not scary for machines to think more and more like humans. What is scary is that people will think like machines. **The equalization of algorithms breaks down boundaries and the walls of intelligence between people, but it also creates contempt, belittlement and domestication of people's individual value. ** How people form their own individual living place is the primary issue. ‍

The following is the transcript of the conversation:

**Q1: After the emergence of ChatGPT, media people have a stronger sense of crisis and worry that they will no longer be useful. What do you think? **

**A1:**In fact, I think news media is an indispensable role in social life. Without this role, there will be major problems in the operating system and decision-making system of the whole society.

**But the value of this role does not lie in itself, but in the position given to it by society. If the value of this role is lost, the normal operating level of society will collapse, and this collapse will cause society to pay a heavy price. **

This is also a kind of social maintenance of the role value of the media.

On the other hand, the major of media itself looks at society from the perspective of communication, media, and news. The role of the media, the value of the media, the logic of the media, and the basis for the views of a certain thing presented by the media are the basis of the media. Look at the world, evaluate the world's profession and background.

The role value of the media is the value in social life, not what it claims to be.

**Q2: In the era of AIGC, how can liberal arts students (such as journalism and cultural communication majors) improve their career attractiveness/how to get high-paying jobs? **

A2: We once did a research.

Among the graduates who enter Internet companies after graduating from first-tier universities in China, graduates majoring in new communication majors have a monthly salary of about 5,000-8,000 yuan. If they work normally, they can rise to 10,000-12,000 yuan a year later; The monthly salary of professional graduates can reach 30,000 yuan, and after one year, the average salary can be increased by about 15%, and after five years, the salary can be raised to about 500,000 yuan/year, which is accompanied by an increase in rank. **

Under such circumstances, we have made a great wish: **Let the salary of graduates of new majors not be inferior to those of graduates of computer science. **

Graduates of traditional journalism and communication education can generally only engage in work related to publicity, planning, or brand communication in Internet companies. Internet companies such as Alibaba, Tencent, Baidu, and Byte are compared to the level of business-to-social communication If you pay attention, the salary will be higher, but most companies will pay more attention to the technical level, and the salary for social communication will be relatively low.

Therefore, we have made a big change in the curriculum. Students need to be motivated by Internet technology, and they need to know what technology does, how to apply it to the market, and how to commercialize technology. We need to cultivate such a design idea for them.

In 2019, we opened a new major called "Internet Product Manager" at Beijing Normal University, teaching students how to use market insights, user insights, and their own understanding of technology to design products or services for the market. When these talents come to Internet companies, their starting salary is more than 30,000 yuan for technical positions.

**Knowing both technology and society, market and design principles, this is an important direction for future newcomers and even liberal arts students to enhance their social value and market value,**so interdisciplinary disciplines are important in current academic research and talents Training is an important direction of transformation.

**Q3: Why are there so many voices in society who worry about their jobs being replaced by AI? **

**A3:**Why do people react heavily to the emergence of generative AI? Here is a question: Do ordinary workers and farmers have a big reaction to this matter?

In fact, they didn’t respond, and they continued to do what they should do. Instead, they used to think that they were mental workers (elites) who belonged to the middle and upper classes in social life and felt threatened. This was a natural result of their unfamiliarity with new technologies. muscle response. **

During the previous industrial revolutions, the revolution of industrialization was more reflected in human manual labor. A large number of physical labor such as shoulder-to-hand and hand-carrying and logging was replaced by machines. Therefore, ordinary workers were full of hatred for machines and went to work one after another. smash the machine.

But in the end, it was still a failure. First, it could not change the trend of historical development. Second, the voice of manual laborers in social management was extremely small, and their natural influence would not be too great.

And this time, AI can replace most of human labor, such as mechanical, repetitive, computational, and collection tasks. Any work that can be programmed, described by algorithms, and requires data support will be replaced. replaced by AI.

This kind of work happens to be what most mental workers do, so they panic. It is precisely because mental workers have many ways to express their voices that their voices can be heard and become a large-scale system.

However, the development of technology and society will not stagnate because of these voices. For these workers, the current priority is to learn how to use AI to improve their work efficiency and increase their bargaining chips.

**Q4: What should be the relationship between media people, creators and technologies like AIGC? Fusion or game? **

**A4: If we use games to understand our relationship with technology, the direction is wrong. **

In the era of mass communication, everyone was instilled with the view that competition is the case. Taking radio and television as an example, if you choose one channel, you cannot watch another channel at the same time. The same is true for newspapers. Most people will not choose two daily newspapers of the same type. A competitive, exclusive competition.

Today is actually the era of digital civilization, which is different from industrial civilization. **Industrial civilization is a fission-style development, while digital civilization is a fusion-style development. **

Why Internet? It is to form new functions and new values through connection and connection. When creating your own value on the Internet, it is not obtained through competition, but to a greater extent, it is formed and constructed through cooperation, integration, and mutual matching.

We always say that behind the technology is the person who masters the technology. The so-called people who play games with technology are actually the winners of traditional rules, rights and interests. It is these people who are playing games with those who master technology. The game between people, not between people and technology.

**Q5: How will generative AI affect the field of news dissemination? **

A5: Judging from the form of communication itself, a set of rules, concepts, rules, requirements and role-playing formed in the period of traditional mass communication is what we often call "journalism professionalism".

However, in the course of the development of the Internet, news professionalism, as more and more people are able to carry out communication decisions, and more and more popular, the right to communicate and the right to speak out have logically been given to everyone, as long as they is willing to speak out.

There is a book called "The Editor-in-Chief Is Dead" (author Chen Xu, former executive editor-in-chief and political and economic columnist of the American "Newsweek" (Newsweek) Chinese magazine), the author uses the editor-in-chief to replace the role of gatekeeper played by the media in the past, and now This gatekeeper role has been broken through by technology, allowing almost everyone to communicate and share with the society without going through the specific gatekeepers of the editor-in-chief. Therefore, the author believes that such a core role of traditional media has gradually been disintegrated and dispersed.

**"Guarding" is actually a set of rules established in the era of mass media. In the era of Internet communication, with the rise of social media, the popularization of social platforms and short video technology has broken through the elite hegemony of discourse expression, greatly It lowers the "threshold" of content production and social expression, making the era of pan-popular communication in which "everyone can become a communicator" a reality. **

Social media gives everyone an equal right to speak, at least formally, everyone can send messages to the society without being screened by others.

However, changes in production methods such as today's intelligent media and AIGC have made communication take on a completely different look.

**Generative AI breaks through the differences in the ability of different groups of people to use and integrate resources, and enables people to have greater improvements in resource mobilization and expression capabilities. **At least in theory, everyone can use a semantic expression and resource mobilization ability above the social average to carry out social content production and dialogue. This is another major boundary breakthrough and a huge endowment for "vulnerable groups" able.

**Q6: What is the most important function of generative AI in the communication field? **

**A6: One of the most important functions of generative AI, I describe it as human enhancement. **

What is human enhancement? In the past, due to differences in talent or acquired efforts, the general public was divided into the general public and the elite. In the past, the elite class controlled the right to speak and the channel to speak out, but what is the state now?

Now, due to the empowerment of generative AI, for example, in terms of translation ability, the results you translate through applications such as ChatGPT are no worse than those translated by foreign language professionals, and sometimes even better.

In this case, even if I am a foreign language blind, I can use ChatGPT to translate the letter and elegant text, then I can communicate with anyone in the world who has a different native language without hindrance.

**This greatly shortens the professional distance between professionals and the general public. **Although this distance may not be completely eliminated, there may still be a gap in some aspects, but at least in terms of fundamentals, their distance is close, even to a negligible level.

**This kind of human augmentation bridging the gap created by the different talents of human beings is especially important, and it will structure the interaction relationship and corresponding pattern of future society. **

**Q7: What are the differences between domestic companies and overseas companies in the direction of research and development of generative AI? **

A7: The reason why Open AI was able to make ChatGPT is mainly because they announced that they are a public welfare investment project without too much pursuit of profit, and there is no profit pressure on the scientists and engineers in the project, so they can Concentrate on solving key technical issues. On this basis, they then carry out commercial realization.

**When domestic companies are doing AI research and development, they will pay more attention to the realization and output of value. Before Baidu Wenxin’s statement was released, hundreds of companies had already announced cooperation and queued up to sign contracts. It has been revealed here that Baidu has clear awareness and requirements for capital realization and value output during R&D. **

Therefore, we will see that domestic large-scale model applications are relatively weak in dealing with fundamental issues, but are extremely capable in the professional field, but the general public rarely experiences professional capabilities, and can only see their general In the question, the answer seems very retarded.

**Q8: Is there a big difference in the level of generative AI products/big models at home and abroad? **

A8: In fact, the ChatGPT made by Open AI is not a major technological breakthrough in the so-called deep learning or development logic of AI. It just matches together such as large models, pre-training, massive data feeding, etc. , When these aspects are reached to the critical point, chemical changes are produced rather than just superposition at the physical level.

In the past, people studied more discriminative AI represented by "recognition-analysis". Now we have explored and made certain achievements in generative AI represented by "synthesis-reconstruction".

Like the element carbon, which can take the form of coal, it can also take the form of diamonds. The same is true for AI from discriminative to generative, which is a phase change (the process of changing a substance from one phase to another, often used to describe the transition of a substance between solid, liquid and gas), not technology A breakthrough in the depth of the vertical field itself.

We also have a lot of talents and experts in AI in China. Although there are certain obstacles in the computing power of chips, what others can do with 20 chips, we can do with 50 chips, but the cost will be higher. will become a bottleneck in research and development.

**But in terms of data volume, because many of our platforms are not open and interoperable with data, they cannot use each other's text materials, and a large number of overseas academic databases and resource libraries are open, so the English corpus will be larger than the Chinese corpus. **

**This will lead to a result that the quality of the answer obtained by searching in English is higher than that in Chinese. **It is the same whether it is an overseas or domestic application.

**Q9: What new changes will AIGC bring to future social life? **

A9: Marx summarized the future society as "a union of free people" more than 100 years ago.

That is to say, the organization method of the future society, the connection method between people is not under the control of an organizational system, workers and tools without the right to choose, but as a subject, it is based on equality If you get along with others, you will get together, if you don't, you will be separated.

There is a new term in the Internet world called "DAO" (Decentralized Autonomous Organization, a brand new way of human organization coordination). This social autonomy method is a bit like what we call circles, but there is still a gap between circles and DAO.

Every member in the circle does not have the right to confirm the data, value and content they create, all of which belong to the platform.

So there are gradual upgrades of web 1.0, web 2.0, and web 3.0. Web 1.0 is what you can see and accept, web 2.0 is what you can use, and web 3.0 is what you can use and own. After the right of web 3.0 is confirmed, the society will become the basic organizational state of the future society like DAO.

Under such a background, in fact, we don’t need to be attached to anything, but to make different combinations of functions and values with different people when completing a certain function and realizing a certain value. .

**This is a pattern of the future society, so in such a society, what we need is not competition, but cooperation and integration of design and creativity. **

Because with a good cooperation, we can do this thing; without this good design, no matter how capable we are, we can't do this kind of thing. Because there are more and more long-chain and multi-element things today, and this thing cannot be done without any one party.

**The first law of the future society is integration, coordination and integration. **

**Q10: What changes can AIGC bring to people's lifestyle? **

**A10: First, it will shorten people's working hours. **

Now the jobs it can replace are mainly those repetitive, normative and mechanical mental work. This is the first wave of replacement, which is about to become a reality.

In another 5 or 10 years, when 50%-60% or even 70%-80% of the total workload of mental workers are replaced by AI, the working hours of humans will be greatly shortened.

Perhaps in the not-too-distant future, we will no longer have an 8-hour working system, but a 4-hour working system. We don't need to work 5 days a week, and 4 days are enough. This frees up a lot of discretionary, non-utilitarian leisure time for us.

**Second, it could widen the digital divide. **

For example, the elderly themselves will have huge learning obstacles in learning new technologies, because the acceptance will indeed be worse when they are older.

But at the same time, it is an opportunity for young people. Without the inertial, structured obsession with the past, the sensitivity to learn new things is much higher.

Therefore, it is an opportunity for young people, but it is a kind of oppression for old people.

**Third, some jobs will be replaced, but there will also be demand for new jobs. **

For example, a very popular post overseas is called "prompt word engineer", which is to help users ask questions to AI, so that users can get high-quality answers. This position mainly depends on the individual's ability. Some people are good at asking questions, and some people do not have such ability. Under such differences and needs, this job was born.

Another type of job is to do things that AI cannot do. AI must have sufficient computing power and data to give an accurate answer. This calculation may take a long time, perhaps decades. But human beings have intuition and understanding, and can give roughly correct directions without complicated calculations.

For example, let AI draw a beauty, and it can combine all the beauties in the world into an average beauty. But such an image, although it does look beautiful, feels that there is some vividness that is seriously lacking. why? There are no defects.

It doesn’t have such ups and downs, because people need to look at it against a somewhat imperfect background to bring out her beauty and vividness. The AI-generated ones are all well-regulated and plain. This kind of work requires sensibility AI can't do it.

** In the future, there will probably be two types of occupations, one is new jobs based on the new AI platform, mainly service jobs; the other is jobs that AI cannot do. **

**Q11: In addition to emerging jobs, will the future job scene be different? **

A11: In the past, human labor was basically done in the fields and in factories. This is the most basic social value production method based on bottom-level labor. Only a few people are in the management class, and most of the laborers either farm the land, raise fish, etc., and they are all in the most basic unit of the front line.

During the industrial revolution, social division of labor was formed. Some people can produce relatively easily with the help of machines, some do scientific research, some do management, and some are environmental managers. In fact, this society is more humanized, and people don't have to be exhausted by that kind of physical labor.

**The same is true in the future. People can use their intelligence to accomplish things that AI cannot. In this process, people's creativity is constantly stimulated, and it may even be like playing a game. **

Because AI is a particularly good assisting service for games. In the past, we could only complete specific tasks in the game according to the rules, and obtain some kind of feedback or effect, so that our body and mind could be released and refreshed.

**Now with generative AI, games can actually be regarded as a social place, a place for social cooperation, and a place for social discussion and innovation. **

In the specific scene of the game, you can communicate with each other, everyone can cooperate, and then innovate and create together, it will become a brand-new execution platform for social combination.

In the scene of game creation, it is much richer than the real scene of society. We can release certain potentials that we have not discovered in different scenes, and if we extend this ability, we can germinate creativity. Creativity may be able to benefit society as a whole and benefit others.

**Q12: Will there be hidden problems and dangers in the social life that coexists with AIGC? **

**A12: The biggest feature of the algorithm is equalization. This kind of equalization breaks the boundary and breaks down the walls of intelligence between people. It seems to be a good thing for the general public, but for the development of human civilization, this rebalance Regression will form a so-called contempt, belittlement and domestication of human individual value. **

Then this will lead to a question: how does a person develop his own personality? **

People are not the result of pure calculations, they are not absolutely rational, they have their own preferences and passions. Like philosophers, good theoretical philosophers stick to their ideas in a paranoid way, not omnipotent.

Therefore, the development of a real person is actually opposite to this average trend. He must emphasize his own preferences, his own personality, and his irrational enthusiasm and emotions.

In the context that algorithms have flooded the entire world, how do people form their own personal living place is a big problem.

**Machines are thinking more and more like people. In fact, there is no need to worry too much. What should be more worried is that people will think like machines. **

**Q13: How to solve the above-mentioned dilemma brought by AI? **

**A13: **In the environment of artificial intelligence in the future, there may be a new way of self-cultivation. **Humans should leave the background environment of all artificial intelligence services and go to a relatively wild and natural scene. Going to stay for a few months and then returning to the artificial intelligence environment can be called scene cultivation and restoration. **

In natural scenes, you need to think about what to do and how to improve your environment. At least let people get out of this environment full of algorithms and intelligent automatic services, and enter the natural world without automation and algorithms, so as to form the recovery of people's basic perception functions and survivability.

** In the future, there may be more powerful anti-intelligence technology and anti-intelligence life scenes. We may only need a week or half a day to stay inside and be able to deal with the AI-imposed effects on you for a period of time. The influence rebounds, so that people maintain a balanced state of rationality, personality and passion. **

**Q14: How to solve the big data infringement problem caused by generative AI? **

A14: Compared with traditional plagiarism, AI data infringement is difficult to be confirmed, because it will disassemble the elements and then integrate them. Maybe the idea is yours, but from the language expression But it's hard to see.

To determine the ownership of intellectual property rights between the designers and contributors of AI works, it is necessary to face technical complexity, which is also an urgent problem to be solved in the future.

**Currently, the more mainstream way is to judge by "digital watermark". **

Every step of using AI to generate works will leave traces of watermarks, such as whether to use AI to collect materials in the stage of material collection, which works are used for reference in the creative stage, and what technology is used in the processing stage, all will generate corresponding watermarks.

When such a work is finally presented to the public, it will be accompanied by various watermarks, and through the background information carried by the watermarks, it can be traced back to various production links.

Such a judgment method requires the assistance of the platform. For the platform, it is very simple to use algorithmic data and technology to judge whether the product is synthetic or original, because the signal carried by the original work is different from that of the synthetic work. Signals are quite different.

This is what the platform should do. For example, in the era of mass communication, when we place an advertisement in a newspaper, we must remind readers that it is an advertisement so that readers do not mistake it for a news report. In the era of AIGC, platforms that publish AI works should also make such an explicit statement.

But this is a theoretical situation. In a real environment, copyright ownership still faces complicated situations. We generate a painting through AI, maybe we have made some imaginary text descriptions to make it generate such a work, but in fact, how much of our real creative part is? This question is difficult to judge.

A wonderful work is inseparable from advanced software, accurate description and human creativity, so how much is the copyright ownership between the creator of the painting and the owner of the AI platform that generated the painting, and how to calculate it? This is another complex issue.

The confirmation of infringement and the attribution of copyright are an answer that requires all parties in the society to compete with each other. It is not possible to sit in an office and check the copyright share of AI works without practice.

Formulating rules on copyright issues needs to balance the interests of all parties, so it also needs to go through a period of relative chaos, which is the cost that must be paid for exploring new things.

**Q15: How should we deal with the “emergence of intelligence” brought about by generative AI? **

**A15: **When ChatGPT just set off an AI boom, the network information department immediately released the "Generative Artificial Intelligence Service Management Measures (Draft for Comment)".

Although there is no position yet, and there is no legislative implementation, but the action is so quick to issue management comments, which in itself shows that the network information department is very sensitive to generative AI and its impact assessment is very high. I hope to be able to compare Precautionary restrictions are placed on the problems it may cause at an early stage.

**For the management of generative AI, the solution is to distinguish which things should be advocated and boosted, and which things It should be balanced and gamed. **

This requires some authoritative professionals to interpret it, government departments to make regulations, and the improvement of public literacy.

When the public sees the negative products generated by AI and consciously does not accept or even dislike them, this is the improvement of public literacy.

The management of generative AI should be a comprehensive result formed by all links. For such a comprehensive result, one platform, one group, or one institution cannot be required to complete it.