In-depth | Obama talks about artificial intelligence again: We need a new social contract to adapt to new technologies

Joi Ito, Scott Dadich, and US President Barack Obama met in the Roosevelt Room of the White House on August 24 this year.

● ● ●

How does artificial intelligence bring us social value?

Scott Dadich: First of all, thank you both for inviting me to the White House. Mr. President, I would like to ask you how you have been doing recently?

Obama: I am busy, but I am doing a lot. I am sure you are aware that there are many political issues that need to be resolved internationally.

Scott Dadich: Next, we will focus on artificial intelligence, a field that has jumped out of science fiction and become a reality, and has a profound impact on our lives. I would like to ask Mr. President, when did you begin to realize the advent of the AI ​​era?

Obama: I know about AI because it's starting to permeate our lives in various forms, and we're still not aware of it. Another part of the reason is that our understanding of AI is mainly derived from popular culture. In fact, there is a difference between broad and specialized AI, and this may be different from what many readers know. In science fiction, all we know is broad AI, right? Computers start to become smarter than us, realize that humans are not always so useful, and even imprison humans in greenhouses, or create a virtual world like a matrix.

In discussions with some scientists, I learned that such scenarios are still a long way off. However, these ideas can free our minds and make us realize that we still have enough freedom to apply specialized artificial intelligence to various important fields, that is, to solve increasingly complex problems through algorithms and computer processing. From medicine to transportation and even the distribution of electricity, we have already witnessed how "specialized" artificial intelligence plays a role in daily life, and it will undoubtedly improve productivity and efficiency at the economic level. If used properly, it can certainly make society prosperous and bring countless opportunities. But the problem with it is that it will reduce jobs, perhaps increase inequality, and reduce wages.

Joi Ito: This question also bothers the students in my MIT lab, but what I am concerned about is that most of the people who have access to these core AI computer technologies are boys, especially white men, and they are obviously much more comfortable talking about computer issues than talking about racial issues. Many of them feel that if they can use artificial intelligence to create scenes like those in science fiction, we don’t have to worry about so many political and social issues. Students will feel that creating such machines is also a way for humans to get rid of these problems.

OBAMA: Yes, that's right.

Joi Ito: But they also foresaw the difficulties that might arise. I think we have come to an era where AI issues go beyond computer issues. It is very important that everyone understands how AI works. In the Media Lab, we use the term "extended intelligence". Because the question we face is, how do we build social value on top of AI?

Obama: Joi used the example of autonomous driving during lunch. Autonomous driving technology is developing rapidly. These machines equipped with autonomous driving technology can not only make decisions quickly on behalf of humans, reduce the occurrence of traffic accidents, and greatly improve traffic efficiency, but also help humans solve carbon emissions and alleviate global warming. But Joi mentioned a crucial point: What is the point of installing these technologies on cars? On the road to autonomous driving, we are also faced with a series of choices: when the car is driving automatically, if an accident occurs, you can guarantee that you will not hit pedestrians, but you may hit a wall or put yourself in a dangerous situation. What should you do if you encounter this dilemma? This becomes a moral and ethical issue. We need some rules to solve these problems. But who should make these rules?

Joi Ito: We once did a "car trolley problem" project (a study by an MIT laboratory on how to solve problems faced by self-driving cars. For example, when faced with a dilemma, should we sacrifice five passengers or five pedestrians?). In the study, we found that most respondents agreed that passengers should be sacrificed to save the lives of others. Some respondents, after being asked this question, said that they would never buy a self-driving car. (Laughs)

● ● ●

What role should the government play in scientific and technological progress?

Scott Dadich: What role should government play in addressing these ethical issues?

Obama: I have been thinking about this issue since the emergence of AI technology. I believe that at the beginning of the development of science and technology, it is a time of "letting a hundred flowers bloom and a hundred schools of thought contend". At this time, the government should fully encourage and support the development of technology, increase investment in research and ensure good communication between basic research and applied research. As technology advances and matures, we should think about the relationship between technology and the existing regulatory system. It is not easy to make this relationship develop in a healthy way, so we face more difficult challenges. At this time, the government should strengthen its connection with the science and technology field. We should first ensure that the government's regulatory system is valuable and can promote the development of technology, rather than forcing new technologies to quickly connect with the government. If we don't do this, the government will become an obstacle to technological progress, which is very detrimental to some people and teams.

Joi Ito: This dilemma reminds me of autism, which is mentioned in the movie Temple Grandin. The heroine said that if Mozart, Einstein and Tesla were alive today, people would think they had autistic tendencies.

Obama: They did grow up in an autistic environment.

Joi Ito: Yes, that's right. If we had the ability to cure autism and make everyone "normal," I think a lot of MIT kids would be different. Whether we're talking about autism or diversity, these options will only appear if we can make them happen. Even so, we don't want a child like Einstein, because although he is extremely smart, he is too eccentric. But without Einstein, our society and science would not have progressed so quickly.

Obama: This gets to a much bigger question, one that has been bothering us since the beginning of AI. A big part of the reason our lives are meaningful is because they are imperfect. People who create art or invent things are often eccentrics, they stand out from the crowd and go against the grain. But it is these qualities that make them talented and drive society forward. If a system is perfect, it must be static. The reason we are what we are today and why we are so vibrant is that we are dynamic, evolving, unique, and surprised by the birth of new things. But in the long run, we always have to adapt to the world. Now the question is, when, where, and in what scenarios can we allow ourselves to accept such changes with a peaceful mind?

Scott Dadich: When AI is compatible with the government, private enterprises and academia, it must have developed to a higher level. At this time, who is the AI ​​research center in the world? Or does such a role exist?

Joi Ito: I think MIT will certainly say that the research center is here (laughs). Historically, the research center may be an academic team supported by the government. But now, most of the laboratories supported by billions of dollars have been commercialized.

Obama: I know the people who support these agencies, and if you ask Larry Page or anyone involved, you can imagine that they would all say in unison that the last thing they want to see is bureaucratic redundancy that limits the development of technology.

Part of the problem we face now is that basic research that serves society has not been well implemented, and ideological obstacles have caused some scientific and technological professionals to lose confidence in government support.

Whenever we talk about a big technological achievement, we always draw an analogy to the first moon landing, and I think it will be the same 50 years from now. Someone once told me that the moon landing cost only 0.5% of the entire GDP, which doesn't sound like much. But today, that number is $80 billion. If the development of AI today is comparable to the moon landing, then we should be investing $80 billion in AI every year, but the actual investment is less than $1 billion per year. There is no doubt that this number will grow. So, the government should strengthen its support. If the government does not support it, then everything we discuss later will be out of the question.

Scott Dadich: Your thoughts make the question that Joi raised even more serious: there is a difference between technology that everyone has and technology that a nation has (like the moon landing). So how can we achieve the transition between these two types of technology so that technology benefits everyone?

Obama: I want to emphasize that the government's support for the development of science and technology does not mean that the government will monopolize this technology or that this technology can only be used in the military. For example, the government supports the development of medicine and allocates funds to build a genetic database for Americans, but every American citizen can use it. Therefore, government-supported technology will be shared with everyone and benefit everyone.

Scott Dadich: But we will face some risks. Musk and Nick Bostrom have also worried about the potential of AI to surpass humans and bring threats to humans. As we move forward, we must not only protect ourselves, but also protect everyone. How can we do this?

Obama: Let me start with something that worries me more. If you have a computer that can play Go, it has powerful algorithms that can help you maximize your profits on the New York Stock Exchange. If someone or an institution can get this technology before others, it will quickly cool down the stock market.

If the algorithm is stolen by a bad actor, it could infiltrate a nuclear program and learn how to launch a nuclear weapon. If the algorithm can learn by itself and is very effective, then we are in danger. I think the NSA should not worry about machines that will replace the world, but about technologies that may infiltrate the national mechanism and threaten human security. In other words, we urgently need to do better now, because the team that damages the country is likely to get ahead of us.

Joi Ito: I basically agree. Some people think that general AI will appear in 10 years. But in-depth research shows that if this kind of AI is to appear, we must have several major breakthroughs in AI technology in these 10 years. So if the breakthrough occurs, you can regulate it.

Obama: So what you need to do is put people in key places (laughs). When you see something like this is about to happen, you just stop it.

Joi Ito: It is important to identify individuals and teams who want to use AI to benefit humanity and support them with practical actions.

Obama: When we used to talk about security and protection, we always thought about military suppression or building walls. Slowly, our understanding of these two words began to change. Part of the reason why cybersecurity has become so difficult is that this threat is invisible, not as tangible as a tank. Now, we need to change our concept of security and adjust our investment policies to protect everyone's security.

Cyber ​​attacks are similar to epidemics, which is what I often worry about. To prevent viruses from entering our bodies, no matter how high the wall is, it will not help. We need to establish a sound public health system, create a health protection mechanism, and develop more effective vaccines. Therefore, when we face the threat of AI, we should take corresponding actions.

Joi Ito: There is another thing that I find interesting. When I was observing microorganisms, I found that good bacteria can be introduced to fight bad bacteria.

Obama: I totally agree. But I still won't let Sunny and Bo lick me, because every time I walk the dog, I still get uncomfortable when I see them biting or chewing something. (Laughter)

Joi Ito: So, we also have to rethink what is clean. Whether it is cybersecurity or national security, we should rethink our response measures. It is very difficult to formulate effective response measures.

Scott Dadich: Will there be a new round of arms race in cybersecurity?

Obama: There is no doubt that international norms, protocols and verification mechanisms for cybersecurity and AI are in their infancy. Right now, people still have distrust of government, which will hinder cybersecurity measures. If some countries in the world see the United States as a superior cyber power, we will say: If you are willing to restrain yourself, so are we. The challenge now is that some mature countries (such as Russia, China and Iran) hold different values ​​and norms than we do. We will treat this as an international issue so that we can better deal with these issues.

● ● ●

Social contracts and economic models must adapt to technology

Joi Ito: We are in a golden age, and everyone is very happy to communicate. If we can ensure that funds and energy are used to support openness and sharing, then everyone can maintain a positive attitude. The world is a big family, and working alone may not achieve the best results.

Obama: Ito is right, and that's why we have meetings every day to attract interested people to join us. I also want to say that although everyone sets a high tone every time we talk about this topic, we must actually consider its impact on the economy. After all, in the eyes of most people, they don't care whether the singularity will come, they are more worried about whether machines will take away their jobs.

I tend to be positive about this issue, because from a historical perspective, humans have upgraded technology countless times, but the final result shows that we did not starve to death, but new technologies created more jobs and improved our quality of life. Of course, some people's concerns are not unfounded. After all, when the AI ​​wave hits, those who can really benefit quickly are those highly skilled people who can use their talents to extend their markets, sales, products and services with the help of machines.

Those with low salaries and low skills may become increasingly redundant; they may keep their jobs but their salaries may be reduced. If we want to achieve a smooth transition, then the whole society should have extensive dialogue to discuss how to deal with the sharp contradictions during this transition period and how to ensure sustained and inclusive economic growth through training. We are becoming more productive, but if left alone, most of the output will still go to a small group at the top. So how do we solve this problem and ensure that everyone can get a dignified income? In addition, how will human art and culture be preserved after AI "rules" everything? Who will take care of the veterans who have dedicated their lives to the country? The social contract we are proud of needs to keep pace with the times to accommodate these new technologies, and our economic model must also change accordingly.

Joi Ito: It is not intuitive to discuss what jobs AI will replace out of thin air. For example, if computers are advanced enough to understand the entire medical system, then diagnosing diseases may be a piece of cake for them, so experienced doctors may be replaced in hospitals instead of nurses and pharmacists, because the former are too expensive. At the same time, some senior positions we are familiar with, such as lawyers and auditors, may also disappear due to the rise of AI. On the contrary, many industries where computers are unable to do well, such as the service industry, will continue to shine. I don't know what the president thinks about the concept of social basic income, but I think society can find new jobs for those who are replaced by machines, just like academics and artists, whose work is no longer directly linked to money. However, if we want to shift to this model, we must change our inherent social perception that people without money are stupid. But the fact is that there are many smart people in academia who are not rich.

Obama: Ito is really sharp. This is why I mentioned earlier that we need to redesign the social contract. Whether the social basic income policy can still adapt to the development of society may have to be discussed in the next 10-20 years. Ito's analysis of the trend of jobs replaced by AI is also reasonable. Advanced jobs may indeed be replaced by computers, but these jobs have one characteristic, that is, they are highly repeatable. In addition, an indisputable fact is: As AI technology is further integrated, our society will become healthier. The links between production and distribution, work and income, etc. will become increasingly weak, as computers do a lot of the work. Faced with this situation, we have to make difficult decisions: teachers have to pay less even though their work is difficult for computers to do. Therefore, teaching people to value their own work and re-examine their income levels is probably an important issue that needs to be discussed now.

We have a lot of work to do. The government is becoming more people-friendly. In the future, the process of filing taxes will be as simple as ordering a pizza or buying a plane ticket. But the reality is that it is not easy to bring the federal government, state governments, and local governments into the 21st century at the same time. Although the government's elites are comparable to those of private companies in terms of capabilities, there is an unbridgeable gap in the overall level of technical equipment. Take me for example. When I first entered the White House Situation Room, I thought I would be able to use the advanced equipment in "Mission Impossible", but the reality surprised me. Even the real-time video footage sent when we were tracking terrorists was so unclear. The White House's technological equipment really needs a major overhaul.

As for the bigger picture, I firmly believe that if we can solve climate change and stop the sea level from rising, then everything else will be fine. I am very optimistic about this, although we still have a long way to go.

In addition, it is also urgent to manage the Internet safely and transparently. Only in this way can the government catch bad guys in time without overstepping its authority. Otherwise, the Internet will become a tool for the government to oppress the public. However, from a technical point of view, such a goal is not easy to achieve. Take encryption as an example. I have had more than one conversation with technicians and the public church, but no one has been able to give an answer that is in the interests of all parties.

Since this is a high-level visit, I would like to add one more point. As a president obsessed with space, I think it is also an important issue to let our next generation enjoy interstellar travel as soon as possible. In this regard, our private enterprises are ahead of the curve. They do not need government funding, but rely on their own crazy ideas to achieve what we dare not think of. The government's thinking is outdated. Every time we mention space flight, our path is still stuck in the Apollo moon landing era. It is indeed time to find new breakthroughs.

● ● ●

Discussion on Star Trek

Scott Dadich: Mr. President, I know you're a big Star Trek fan. But to be honest, the show is just a technological cover, but the core is still utopian. How has it shaped your vision of the future?

Obama: I was a die-hard fan of Star Trek when I was a kid. The show was so addictive. In fact, what made Star Trek so popular was not the technological elements, but the values ​​and friendship. So even if the show used cheap special effects, we still enjoyed watching it. It taught us what human nature is and gave us the confidence to solve problems.

In my opinion, The Matrix is ​​a true legacy of Star Trek. It's not that the grand and complex setting of the movie moved me. On the contrary, it shows us the attitude that many people should have when doing their work seriously. With this attitude, we will move forward courageously, and this is also the spiritual core of the American dream that I have always emphasized. Whether we succeed at once or not, we will boldly try and reach the top by stepping on our own experience of failure. If we lose this spiritual core, then everything we rely on will collapse in an instant.

Joi Ito: Mr. President, what you said is great. I also like the optimistic spirit conveyed by Star Trek. However, I still advocate letting a hundred flowers bloom. I think the villains in the movie are still a bit stereotyped. They may not be bad. Some of them are just misled.

Obama: Star Trek, like any good story, helps us realize that we are complex creatures, with both light and dark sides. The boldness I mentioned earlier fits this theme perfectly, without which we will certainly not be able to eliminate barriers and divisions. There is always a conflict between rationality and humanity, and this is also true with AI. Although we talk a lot about the future of AI, humans have only just scratched the surface of this technology. So we must remain in awe and boldly experiment.

via wired