The State Duma has proposed steps to reduce the risks of using AI
Artificial intelligence (AI) technologies certainly expand the possibilities of designing the future, but they carry certain risks, and they are quite obvious. Alexander Asafov, a State Duma deputy and member of the Committee on Science and Higher Education, told Izvestia on October 29 during a round table discussion at the Expert Institute for Social Research (EISR) following the conference "Social Architecture: Theory and Practice of Social Change."
"These are certainly new opportunities. This includes information analysis, scenario development, explanation of complex concepts, and scientific research. Take a look, for example, at what happens in mathematics in this part. Gloomy colors <...> were certainly also present, because the risks also exist, these risks are obvious," the deputy believes.
The parliamentarian noted that setting tasks for artificial intelligence should be based on trust in people who are able to criticize these tasks.
"If the interlocutor can point out a mistake in order to correct it together, then artificial intelligence has a certain obsequiousness. Of course, this is hardwired into his commercial model, and he will try to find the very answers that the user is ready to harness without checking. And here the second reason is human verification of human results," Asafov noted.
Thus, people will receive a certain product as a result of their joint work with generative neural networks. However, the decision after checking it remains only for the person. According to Asafov, the fourth task is to take responsibility for the results of AI work. Speaking of designing the future, an algorithm cannot be responsible for what happens as a result of neural networks.
"Therefore, no matter how perfect AI is, in my opinion, it is inappropriate here, it is important to take this into account. After all, there are such features as human perception, which, in fact, shifts this responsibility to algorithms and, in general, to the machine," Asafov summed up.
Russian President Vladimir Putin said on October 1 that there is a danger of digital totalitarianism in the world due to the emergence of AI technologies. He noted that experts are increasingly saying with alarm that the spread of AI can lead everyone to depersonalization and a common denominator by using externally customized algorithms. Therefore, "digital totalitarianism" would become the same dead-end, gloomy and inhumane path as the totalitarian societies of the past.
Переведено сервисом «Яндекс Переводчик»