New Nobel Prize winner Acemoglu: How do you view the current development and risks of AI?
Original author: Chen Qinhan, reporter of The Paper
On October 14th local time, the Royal Swedish Academy of Sciences announced that the 2024 Nobel Prize in Economics will be awarded to Daron Acemoglu, Simon Johnson and James A. Robinson in recognition of their "research on how institutions are formed and their impact on prosperity."
The three economists demonstrated the importance of social institutions to a country's prosperity, the judging panel said in a press release. "Societies with weak rule of law and institutions that exploit their populations cannot generate growth or bring about positive change, and their research helps us understand why."
Acemoglu was born in Istanbul, Turkey in 1967. He has been teaching at MIT since 1993 and won the Clark Prize in 2005. His research covers a wide range of areas, including political economy, economic development, economic growth, technological change, inequality, labor economics and network economics. He has co-authored a large number of papers with the other two economists who won the award this time, and co-authored bestsellers such as Why Nations Fail and The Narrow Corridor with Robinson.
In recent years, one of Acemoglu’s research focuses has been on the impact of automation technologies such as industrial robots on the labor market. In 2023, he co-authored the book “The Impact of Automation on Labor Markets” with Simon Johnson.Xiaobai NavigationPower and Progress is published, which discussesAIThe dilemma faced by the most important technology of our time.
"Much of my research focuses on the interaction between political economics and technological change, which are two major forces that shape our capabilities and growth opportunities, while also influencing our political and economic choices," Acemoglu said in an exclusive interview with The Paper in June this year.
His research found thatcurrentAIThe trajectory of development is repeating and exacerbating some of the worst technological mistakes of the past few decades.For example, there is too much emphasis on automation without sufficient investment in creating new tasks. He believes that business leaders need to realize that their greatest asset is their workers, and instead of focusing on cutting costs, they should look for ways to increase workers' productivity, capabilities and impact.
阿西莫格鲁非常担忧AI成为将财富和权力从普通人转移到一小群科技企业家的方式，为打破大型科技公司的政治力量，他认为“Antitrust alone is not enough; we need to redirect technology in a direction that is beneficial to society.”.
He suggested that three principles could be considered to guide the development of AI: first, prioritize the usefulness of machines; second, empower workers and citizens instead of trying to manipulate them; and third, introduce a better regulatory framework to hold technology companies accountable.
*The following is an exclusive interview with Acemoglu published by The Paper on June 16, 2024. The original title is《Exclusive Interview | MIT Professor: Worried that AI will become a tool to transfer wealth and power to a few technology entrepreneurs》. ”
Reference News Network quoted CNN on June 15 as reporting that Apple surpassed Microsoft on the 13th to become the most valuable listed company in the United States. Last week, the company announced a series of news including the generative artificial intelligence function of the iPhone at its annual Worldwide Developers Conference, and its stock price has been rising.
Apple, Nvidia, and Microsoft have been competing fiercely for the title of the world's most valuable company. After Apple redefined "AI" - Apple Intelligence, its market value surpassed Nvidia, whose value skyrocketed due to its AI chips, and then surpassed Microsoft to regain the top spot. Apple's current market value is $3.29 trillion, slightly higher than Microsoft's $3.28 trillion. Generative artificial intelligence has become the core driving force behind the rise in the market value of the three major technology giants.
In the face of this AI craze, the National Bureau of Economic Research recently released a paper written by Daron Acemoglu, a professor at MIT, pointing out that the productivity improvement brought about by future advances in artificial intelligence (AI) may not be large, and it is estimated that the upper limit of AI's growth in total factor productivity (TFP) in the next decade will not exceed 0.66%.
Acemoglu points out in the article that generative AI is a promising technology, but unless the industry undergoes a fundamental repositioning, including major changes to the architecture of generative AI models (such as large language models LLMs) to focus on reliable information to improve the marginal productivity of workers in various industries, rather than prioritizing the development of general, human-like conversational tools.
Acemoglu, a Turkish-born American economist known for his work in political economy and the interplay between political economy and technological change, is skeptical of some of the overly optimistic predictions about AI’s impact on productivity and economic growth.
Last year, he co-authored a new book, Power and Progress, with British-American economist Simon Johnson, which talked about the AI revolution that could subvert human society. They believe that the current development of AI has gone astray, and many algorithms are designed to replace humans as much as possible. "But the way to make technological progress is to make machines useful to humans, not to replace humans."
Mira Murati, chief technology officer of Open AI, said at an event in May that they are not only focused on enhancing the functionality and usefulness of the model, but also on ensuring that it isSafety, so that it is consistent with human values and does not get out of control, thereby creating AGI that benefits mankind.
"The more I study the capabilities and direction of AI, the more I am convinced that its current trajectory is repeating and exacerbating some of the worst technical mistakes of the past few decades." Professor Acemoglu recently told The Paper (www.thepaper.cn) that most of the leading players in the field of AI are driven by an unrealistic and dangerous dream, namely the dream of realizing general artificial intelligence, "which puts machines and algorithms above humans."
Some analysts regard Acemoglu as an AI pessimist. In response to The Paper, he said that as a social scientist, he would pay more attention to some negative social impacts.
Acemoglu often works with his wife, Professor Asu Ozdaglar, who is the head of the Department of Electrical Engineering and Computer Science at MIT. Although they are in different fields, the couple's views on the development of AI are consistent in the general direction, but Acemoglu also admits that his ideas may be more pessimistic than his wife's.
As the race for commercialization of artificial intelligence accelerates, AI big models are competing fiercely, but there is no doubt that technology giants such as Open AI, Microsoft, Google, and Nvidia have already seized the opportunity in the development of AI. Acemoglu said that he is very worried that AI will become a way to transfer wealth and power from ordinary people to a small group of technology entrepreneurs, and the "inequality" we see now is the "canary in the coal mine."
Technology and society: The greatest asset is people
Q1:Your research covers political economy, technological change, inequality and other fields. Under what background and opportunity did you begin to pay attention to the role of technological development in inequality? What was your initial view on technological development, and how did it evolve into your current advocacy that "the current development path of artificial intelligence is not conducive to the economy or democracy"?
Acemoglu:
Much of my research focuses on the interaction between political economy and technological change, two forces that shape our capabilities and opportunities for growth while also influencing our political and economic choices.
AI has become the most important technology of our time, partly because it has attracted a lot of attention and investment, partly because it has made some impressive progress, especially with the improvement of GPU performance, and partly because of its ubiquitous impact. These factors prompted me to conduct research in this field.
The more I study AI’s capabilities and where it’s headed, the more I’m convinced that its current trajectory is repeating and exacerbating some of the worst technological mistakes of the past few decades:Overemphasis on automation, just as we prioritize automation and other digital technologies without adequately investing in the creation of new tasks; andSocial platforms try to profit from people’s data and interests, and therefore all the mistakes made.
I am also particularly concerned about the fact that most of the top players in the field of AI are driven by an unrealistic and dangerous dream of achieving general artificial intelligence, which is a way for these top players to dominate everyone else, by placing machines and algorithms above humans.
Q2:Advanced computer technology and the Internet have enabled many wealthy people to transfer wealth and made technology giants more powerful than ever before. Despite this, we still accept such technological innovation because it also brings positive effects. Technological change has both advantages and disadvantages, and history shows that society always finds ways to adapt to new technologies. With a new wave of technology sweeping in, why do you think the issue of inequality is particularly worrying?
Acemoglu:
I agree with the above when it comes to social platforms and artificial intelligence, but I have a different opinion when it comes to the Internet. I think the Internet is used in some ways inappropriately, but of course I do not deny that the Internet is a very beneficial technology and has played a very important role in connecting people, providing information to people, and creating new services and platforms.
I am very concerned about AI as a way to transfer wealth and power from ordinary people to a small group of tech entrepreneurs.The problem is that we don’t have any of the control mechanisms necessary to ensure that ordinary people benefit from AI, such as strong regulation, worker involvement, civil society, and democratic oversight. The “inequalities” we see are the “canaries in the coal mine” that mean worse things are coming.
Q3: You point out that inequality caused by automation is "the result of how companies and society choose to use technology." As the market power and influence of technology giants grow and may even get out of control, what is the key to our response? If you were the CEO of a large technology company, how would you use AI to manage the company?
Acemoglu:
My advice to CEOs is to realize that their greatest asset is their workers.With its specialInstead of focusing on cutting costs, we should look for ways to increase worker productivity, capabilities, and impact.This means using new technologies to create new tasks and develop new capabilities for workers.
Of course, automation is beneficial, and we will certainly use more of it in the future, but it is not the only thing we can do to improve productivity.Automation shouldn’t be the only thing CEOs pursue and prioritize.
Q4:US antitrust enforcers have publicly expressed a series of concerns about artificial intelligence, and the US Department of Justice and the Federal Trade Commission have reportedly reached an agreement that paves the way for antitrust investigations into Microsoft, OpenAI, and Nvidia. Can such antitrust actions against large technology companies really increase market competition and prevent AI development from being dominated by a few companies?
Acemoglu:
Absolutely. Antitrust is important. The root of some of the problems in the tech industry is the lack of antitrust enforcement in the U.S. The five largest tech companies have all established strong monopolies in their fields because they have been able to acquire potential competitors without any regulation. In some cases, they have purchased and disabled technologies that could compete with them in order to consolidate their monopoly. We absolutely need antitrust to break up the political power of big tech companies, which has become very strong over the past three decades.
But I also want to emphasize thatAntitrust alone is not enough, we need to redirect technology in a direction that is beneficial to society. It is impossible to achieve (increasing market competition and avoiding a few companies dominating AI development) by simply splitting Meta into Facebook, Instagram, and WhatsApp. In the field of AI, if there is a concern that AI technology will be used for manipulation, surveillance, or other malicious purposes, antitrust itself will not be the solution, and antitrust must be combined with a broader regulatory agenda.
Technology and people: How to avoid repeating the same mistakes
Q5:You have always emphasized "machine usefulness," which is "trying to make machines more beneficial to humans." How do you think this goal should be achieved? What are the consequences of not achieving this goal?
Acemoglu:
This ties into the advice to CEOs above. What we want are machines that extend human capabilities, and with AI, there is a great deal of potential to do just that. AI is an information technology, so we should be thinking about what AI tools can provide useful, context-dependent, immediate information to human decision makers, and what AI tools can be used to make humans better problem solvers, able to perform more complex tasks. This is not just for creative workers, academics, or journalists, but for blue-collar workers, electricians, plumbers, health care workers, and all sorts of other professions.Better access to information drives smarter decisions and higher-level tasks, which is what machine usefulness is all about..
Q6:You suggest giving fair tax treatment to workers’ labor. Are taxing equipment and software like human employees, or reforming taxes to encourage employment rather than automation, realistic solutions?
Acemoglu:
Yes, I argue with Simon Johnson in Power and Progress that a fairer tax system could be part of the solution. In the United States, when businesses employ workers, they face a marginal tax rate of more than 30%. When they use computer equipment or other machinery to perform the same task, the tax rate is less than 5%, which provides excessive incentives for automation while discouraging employment and investment in training and human capital.It is a sound policy idea to bring marginal tax rates on capital and labor to the same level.
Q7:You have proposed tax reforms to reward employment over automation. How would such reforms affect companies’ adoption and investment in automation technology?
Acemoglu:
We have to be careful not to discourage investment, especially in a world where many countries need to grow rapidly and need new investment in areas like renewable energy and healthcare technology. But if we can encourage technology to develop in the right way, it will be good for business. So my proposal isRemove excessive incentives for automation and hope it can be done in a way that doesn’t generally discourage business investment.
Q8:The rapid development of social platforms has brought some negative effects, such as information bubbles and the spread of misinformation. How do you think we can avoid repeating the same mistakes in the further development of artificial intelligence?
Acemoglu:
Three principles can help avoid repeating the same mistakes: (1) prioritize machine usefulness, as I advocate; (2) empower workers and citizens rather than try to manipulate them; and (3) introduce a better regulatory framework that holds tech companies accountable.
Technology and Industry: Digital advertising tax makes the industry more competitive
Q9:Technology expert Jaron Lanier emphasized the issue of data ownership of Internet users. How do you think the ownership and control of personal data should be better protected in terms of policy?
Acemoglu:
I think this is an important direction. First, we will need more and more high-quality data, and the best way to produce this data is by rewarding people who create high-quality data, and the data market can achieve this. Second, data is currently being plundered by technology companies, which is neither fair nor efficient.
However, the key point is that the data market is not like the fruit market. My data is often highly substitutable for your data, so if technology companies can negotiate with individuals to buy their data, there will be a "race to the bottom", and the administrative costs of doing so will be very high. So I think,Well-functioning data markets require some form of collective data ownership, either in the form of data unions or data industry associations, or other collective organizations..
Q10:What do you think about introducing a digital advertising tax to limit the profitability of algorithmically driven misinformation? What impact might such a tax policy have on the digital advertising industry and the spread of information?
Acemoglu:
I support a digital advertising tax because business models based on digital advertising are extremely manipulative and they work in concert with strategies that create emotional rage, digital addiction, extreme jealousy, and information cocoons. They also work in concert with business models that exploit personal data, leading to negative consequences such as mental health problems, social polarization, and a decline in democratic citizenry.
Worse, if we want to re-direct AI as I suggest, we need to introduce new business models and new platforms, but today's business models based on digital advertising make this impossible. You can't start a new social platform based on user subscriptions, and you can't replicate the success of Wikipedia because you are against a company that provides free services and has a large customer base. So,I see a digital ad tax as a way to make the tech industry more competitive: If the "low-level means" of obtaining user data and making profits through digital advertising can be curbed, new business models and more diversified products will emerge.
Q11:Can you share some of the positive changes you think future technological developments may bring about, and how we should prepare for and drive these changes?
Acemoglu:
If we use AI correctly, it can improve the professional skills of workers in all walks of life, and it can also improve the process of scientific discovery. I also think there are ways to use AI democratically.
The article comes from the Internet:New Nobel Prize winner Acemoglu: How do you view the current development and risks of AI?
In the past 24 hours, many new popular currencies and topics have appeared in the market, and they may be the next opportunity to make money. Summary The US dollar index continued to fall, Powell issued a strong signal of interest rate cuts, and BTC remained volatile after rising on Friday. The arrest of Telegram's founder in France caused a sharp drop in TON ecosystem TVL, and TON's coin price rebounded after a sharp drop...
相关文章
