The creation of an ethical environment for the development of these technologies should be based on "human-centrism", believes the Russian scientist Maxim Fedorov.
Participants of the international expert group of UNESCO, which develops ethical standards for the field of artificial intelligence (AI) technologies, during the discussions, came to a common opinion that monopolization is unacceptable in this area. This was told to TASS by the representative of the Russian Federation in the group, the director of the Center for Scientific and Engineering Computing Technologies for Tasks with Large Data Sets of the Skolkovo Institute of Science and Technology Maxim Fedorov.
A special expert group (UNESCO's Ad Hoc Expert Group, AHEG), formed following the results of the 40th session of the General Conference of UNESCO, which, together with representatives of the United States, Canada, Great Britain, EU countries, China and other countries, included a Russian scientist, began in April to work on resolving possible ethical controversies in the field of artificial intelligence. As a result, specialists working in the mode of conference communication (about 20 experts) should form the basis of a UNESCO normative document with recommendations in this area.
“The point is that artificial intelligence should not have a single owner. There should be no monopoly - neither for the country, nor for the company, nor for individuals. This is correct in a number of aspects, and now there is agreement on this issue in the UNESCO expert group. However, in this area, far from everything is simple, the question of control immediately arises and an inflection in the other direction may occur, Fedorov said.
He also said that now at the UN platform for education, science and culture, they are looking for answers to the questions of what to do with de facto technological monopolies already existing in the world and what should be the mechanisms of protection against monopolization of the artificial intelligence sphere. On these and other issues, the position of Russian representatives and a number of other countries is based on the principles of the unconditional priority of human interests, security and freedom of choice, he added.
AI as a member of society
During the discussion of a single ethical concept for artificial intelligence technologies, a number of experts, referring to the approaches of some East Asian countries, came up with a proposal to endow artificial intelligence with the status of a "quasi-member of society", like a legal entity, with its rights and obligations.
Promotional video:
“This topic is very dangerous. That is, AI is created as a kind of new entity that lives by itself, but it has rights - as a member of society. We discussed for a long time what reason, consciousness and intelligence are in this context, but nevertheless we managed to turn the discussion to the fact that the subject of society can be someone who can suffer: a person can, but an algorithm cannot. These thoughts are also traced in the works of visionary Yuval Noah Harari, "Fedorov said, adding that together with Russia, a number of experts opposed the creation of a" quasi-member of society "with references to already adopted regulatory documents in the field of AI, including within the EU.
The creation of an ethical environment for the development of artificial intelligence technologies should be based on “human centrism”, in which the creation of certain types of artificial or symbiotic intelligence would be strictly prohibited, like cloning people, the Russian scientist is convinced.
Responsibility
Debates also suggest that robots and algorithms be given “personal responsibility” so that the machine or system that makes a mistake can be “punished”.
“The trick here is that the one who sold the robot made a profit, the one who uses it, in fact, also benefits. And if the robot makes a mistake that entails victims and losses, then it seems that no one is to blame - let's punish the robot. But someone made and bought it. It turns out that manufacturers do not want to take responsibility and actually shift it to the system users,”the expert continued.
For example, when using a purchased program, the user will not be able to request technical support in case of errors. Instead, the manufacturer will simply remove the defective product from the disc as a "punishment," the TASS source explains.
Access to AI - a duty or a right?
Another topic of discussion is the ability of people to access the artificial intelligence infrastructure. Some experts propose to grant the right to use such systems to every person on Earth. However, Maxim Fedorov sees in this a threat of imposing technology on people against their will.
“Now my colleagues and I seem to have managed to come to a common opinion that life choice (life choice, English - TASS note) is the business of each individual person. We must respect the rights of people, who, by the way, are not obliged to explain why they do not want to use algorithms,”Fedorov said, mentioning as an example the Amish representatives of the Protestant followers of Jacob Amman in the United States who do not use gadgets, the Internet and even electricity.
“I believe that people should have the right to access technology, but they should also have the right to choose and the right to be full members of society without using these technologies. Why? Someone - due to age, someone due to religious beliefs, someone due to experience and other circumstances. But this is a very delicate question, because we are maneuvering between accessibility, which developing countries may not have, and “hook” all people on a certain product in other countries,”he added.
Impact on the environment
The participants in the discussion also consider the prospects for the development of artificial intelligence technologies and the impact of this process on the environment, taking into account the increasing energy costs and volumes of infrastructure components. The representative of Russia did not rule out the likelihood of signing in the future an analogue of the Kyoto Protocol at the UNESCO site, which would limit the negative impact on the environment.
“To what extent this will affect Russia is an ambiguous question. In terms of infrastructure development, our country does not yet reach the world average level of equipment, but the world is changing. In principle, we have everything in order with energy carriers, there are many “cold” territories for placing supercomputers and data processing centers,”the expert explained.
A normative document based on expert findings is expected to be adopted by UNESCO in 2021.