Artificial intelligence governance is facing challenges
May 31, 2024 15:46 Source: "China Social Sciences" May 31, 2024 Issue 2903 Author: Weng Rong

  

 

 

    Artificial Intelligence Law is the current international legislative hotspot。March 13, 2024,The European Parliament officially passed the Artificial Intelligence Act,This is the world's first comprehensive regulatory regulations in the field of artificial intelligence,It will bring an important impact on companies dedicated to the development of artificial intelligence systems。Around artificial intelligence -related governance challenges,Our reporter interviewed Anupam Chander, a professor at the George City University of the United States (also translated "George Dun University") Law School. Roman yampolskiy)。

  bet365 best casino games

"China Social Sciences": How do you think you should prevent risks related to artificial intelligence,For example, invasion of privacy and algorithm prejudice?

Qiande: There are currently two types of risks。One is a common risk that people are familiar with,For example, false information、Direct harassment or directional marketing。Another is the risk that people are not familiar with。For example,Artificial intelligence may develop self -consciousness and resist its controller。In some people's eyes,Worrying about the second risk is unnecessary,They think this kind of concern is pure speculation and interference items。But I think,You should pay attention to these two risks at the same time,It is necessary to ensure the use of existing laws and regulations to fight harassment、Attack or manipulating behavior,and ensure that privacy can be protected when using algorithms、Reduce prejudice,Help artificial intelligence make non -discriminatory decisions。This also means,Must require developers to avoid the above risks as much as possible、Predict the adverse use scenarios、Do a good job of system settings to detect poor behavior and take corrective measures。We need to establish a control system when we foresee related risks,Prepare for targeted manner。

I call the second risk as a sexual risk,This risk is very difficult bet365 Play online games to deal with,It is related to what extent it wants to develop to the artificial intelligence system。Of course,It is difficult to ask technology companies to slow down the pace of development,However, it may be possible to consider requesting enterprises to set up emergency stop switches or carefully consider how to restrain the free will of artificial intelligence systems。People must also consider the impact of artificial intelligence on the employment market,This is a big problem。Because artificial intelligence in certain areas can do better than people,So we need to consider whether the demand for such practitioners will be reduced。Even more,If the decision -making ability of artificial intelligence becomes as powerful as humans,So will the demand for decision makers decrease? These are extremely serious issues that people must think,We need to prepare for technological progress。

"China Social Sciences": Why do you think artificial intelligence and super artificial intelligence will exceed the scope of human control? Will artificial intelligence threaten human survival?

Yangpolsky: If the artificial intelligence system decides to take over the world,Humans will be incomparable。Compared with humans,Faster thinking of artificial intelligence,better memory,It can even solve some problems that humans cannot solve。In my opinion,It is unlikely to control super artificial intelligence。But no one knows whether artificial intelligence will threaten human survival and when will this threat appear。We don't know if this situation will happen。

"China Social Sciences": How do you think we should understand the survival threat mentioned earlier,Does it mean that human beings will be eliminated by artificial intelligence?

Yangpocuski: As far as the threat of survival here,Human beings will be eliminated by artificial intelligence is a possibility that is easy to be understood by the public,But there are other terrible cases。For example,Humans have society、The ability to create culture is destroyed。It's like animals in the zoo,Although not killed,But they may not have a life they want。

"China Social Sciences": Some people think,Artificial intelligence reduces the difficulty of cyber crime,making it more difficult to be detected。What measures bet365 Play online games do you think should be taken to fight against the evolutionary threat?

Qiande: Some people use artificial intelligence to develop malware,​​But some people use artificial intelligence to fight malware、Strive to prevent and prevent the behavior of malicious use of artificial intelligence。This is a competition between good and evil,It is also a problem that exists from the beginning of the Internet。History,Human beings have achieved both success in combating cyber crimes,Also failed。We must not only make full use of existing laws and regulations to combat relevant criminal activities,And need to encourage countries to join、Adopt and execute relevant rules,Ensure that the law enforcement capabilities of various countries。other,It is also necessary to identify hackers at the international level、Campaign to crack down on hackers。

Super artificial intelligence development should not be too fast

"China Social Sciences": How do you think you should prevent the phenomenon of "competition" in the field of artificial intelligence,and make the relevant parties from reducing the security and ethical standards of artificial intelligence in order to pursue economic benefits?

Qiande: Whether artificial intelligence is safe is about interests of various countries。Even if all countries are competing for artificial intelligence,But they also noticed the various risks that artificial intelligence may bring,and are looking for a way to deal with。but forced by competition pressure,There will be a upper limit of the strict level of regulatory artificial intelligence in various countries。I don’t think that there will be a country to slow down the pace of developing artificial intelligence in order to examine risk issues。Artificial intelligence has become a hot spot in global competition,Generally speaking,Global political issues at the global level are difficult to find the optimal solution,This problem is not easy to solve。

"China Social Sciences": Does this mean that we should give up to control artificial intelligence,Or is there any other way out?

Yangpocuski: We can choose not to develop super artificial intelligence,Only developed narrow artificial intelligence that solves specific problems,That is weak artificial intelligence。This is an artificial intelligence suitable for specific applications or tasks,Programming for a single Bet365 app download task,Based on limited parameters、constraints and contexts to simulate human behavior。For example,Facial recognition、voice and language recognition and artificial intelligence program used in autonomous vehicles。These are relatively safe。People understand their intentions,I also probably know how to control these systems。and general artificial intelligence or strong artificial intelligence、Deep artificial intelligence evolved into super artificial intelligence will be fast。,Existing standard test、Monitoring and management methods will no longer be applicable。So,Before the establishment of an effective safety mechanism,I suggest delay the study of all general artificial intelligence systems。

"China Social Sciences": Do you think this can be realized? After all, many companies are already developing common artificial intelligence。

Yangpocuski: I just think it would be better to do this,But it is unlikely that related companies are unlikely to stop development activities。

  Strengthen international cooperation of artificial intelligence governance

"China Social Sciences": The European Parliament recently passed the "Artificial Intelligence Act",Will this produce the "Brussels Effect",That may be aligned with EU standards with EU standards?

Yangpocuski: This is very likely。But I have an attitude that is not very optimistic about solving security problems through the law。The EU's "Artificial Intelligence Act" does not directly and fully deal with the problems of the advanced system,It mainly focuses on privacy involved in existing systems、Bad、Deep fake and other questions,It has little relationship with the survival threat that the super artificial intelligence system and the human survival threat。

"China Social Sciences": How do you think we should design a moral framework or governance model,To effectively monitor the development and deployment of the super artificial intelligence system?

Yangpocuski: This is a complicated question。People have spent thousands of years trying to formulate moral standards that everyone recognizes,But looking around the world,We will find that it is precisely because it has not formulated a good behavior framework,So many battles still exist。This is also one of the reasons for my skepticism of human beings to control artificial intelligence。If we human bet365 live casino games beings cannot reach an agreement,Can't follow ethical and morals to act,That and one with us in architecture、Value and other aspects are so different from artificial intelligence.。

In my opinion,It should be issued a broader global artificial intelligence governance framework。For example,Can countries be able to cooperate and reach a consensus,Perform the suspension order agreed by all parties。But at least I haven't seen this universal consensus at least at present。Some people think this is not a problem at all,It should be developed as soon as possible。I just published a paper on monitoring the restrictions on the operation of artificial intelligence training,Welcome to more scholars to test the results I discussed in the text、Find the error in the text。I will be very willing to know that I am wrong,Knowing that some of the facts I don't think can do it are actually possible。

"Journal of China Social Sciences": Differences in legal and cultural backgrounds may affect the way of artificial intelligence in various countries,How do you think you should consider these differences while effective management of artificial intelligence use?

Qiande: First of all, think about what the specific differences between countries are,Maybe some countries are particularly sensitive to some kind of discrimination,and other countries do not exist in this area。For example,India may pay more attention to the surname discrimination,The United States may pay more attention to racial discrimination,Gender discrimination may be a global problem。So,I think every country needs to ensure that its artificial intelligence use rules are consistent with the laws and regulations in human rights。other,Each country has its own cultural specifications,These specifications should be embedded in the law,and the artificial intelligence system must also abide by these specifications。As for operating problems,My preliminary suggestion is,Different laws of various countries,So technology companies that develop global business need to comply with different rules,Even if the cost is higher、More boring。

"China Social Sciences": You predict in the next ten years,What are the key development of international law in the field of artificial bet365 best casino games intelligence? How should legislators deal with future challenges?

Yangpocuski: Some legislators have tried to limit the development of cutting -edge systems through regulations,For example, limit calculation or training ability,But I think it is unlikely to succeed。Different judicial areas have different regulations,EU passed a certain regulations,Do not mean Britain、The United States will introduce the same regulations,There is always a place in the world that allows people to carry out research unlimitedly。From the historical perspective,Human beings specify many behaviors as illegal。For example,Sending spam is illegal,It is also illegal to make the use of computer viruses and extortion software。But it is illegal to announce that these behaviors have not solved the problem,Questions still exist。As far as network security is concerned,The attacker has a chance to try repeatedly。After a company becomes a victim of extortion software, you can get the opportunity to restore normal business by paying the ransom。But in the face of human survival crisis,We have no second chance,We can't bear any accident。

Journal of China Social Sciences: In terms of the gap between the development of developed countries and developing countries in the field of artificial intelligence,What role do you think the United Nations and other international organizations can play?

What role do you think the United Nations and other international organizations can play?,In addition, they generally face other problems that are more urgent to solve,Therefore, there may be no condition to supervise artificial intelligence。I think the United Nations can do a lot of things。For example,Provide scientific and technological help for developing countries,and helping countries reach an agreement on artificial intelligence legislation。Although I don’t think that countries will really formulate a unified rule,But in terms of dealing with human survival threats,The United Nations can play a significant role。

Editor in charge: Cui Bohan
QR code icons 2.jpg
Key recommendation
The latest article
Graphics
bet365 live casino games

Friendship link:

Website filing number: Jinggong.com Anmi 11010502030146 Ministry of Industry and Information Technology:

All rights reserved by China Social Sciences Magazine shall not be reprinted and used without permission

General Editor Email: zzszbj@126.com This website contact information: 010-85886809 Address: Building 11-12, Building 1, Building 1, No. 15, Guanghua Road, Chaoyang District, Beijing: 100026