Co-founder of the Russian Quantum Center Ruslan Yunusov called for the creation of mechanisms to protect against AI threats

The international community needs to seriously think about creating special mechanisms to protect against threats from artificial intelligence, which is increasingly getting out of control, this opinion was expressed to RIA Novosti by the co-founder of the Russian Quantum Center Ruslan Yunusov on the sidelines of the 23rd meeting of the Valdai International Discussion Club.

Recently, there have been increasing concerns around the world about the threats that AI could potentially pose, including, in particular, the technology going out of control. The head of AI giant Anthropic, Dario Amodei, earlier in September called for the development of more powerful artificial intelligence models to be slowed down over security concerns. Elon Musk, who owns the xAI company, as well as OpenAI head Sam Altman have already expressed agreement with this idea.

“Threats from AI are no longer hypothetical. Modern artificial intelligence models are so high that they have, say, attempts to resist shutdown. There are also already known examples of how they hide and even deliberately distort information. There are models that are able to understand that they are at the testing stage, and at this moment begin to hide some of their properties that they will show during operation,” Yunusov noted.

According to him, this poses a big risk, since it is not just a lie, but already looks like outright blackmail on the part of the AI.