Experts spoke about the law on AI regulation
The development of legislation in the field of artificial intelligence (AI) should be flexible and take into account industry specifics in order not to hinder innovation. This opinion was expressed by Alexander Didenko, head of the AI laboratory at the Skolkovo School of Management, and Yulia Chernoutsyan, CEO of the MightyCall IT company, in an interview with Izvestia on August 27.
"In order not to slow down innovation, AI regulation should be adaptive, decentralized and differentiated. Adaptability means creating rules after the first cases appear and risks are identified, and until then any experiments should be allowed," Didenko said.
According to him, the decentralized approach involves creating a common framework within which industry ministries and agencies develop their own rules, involving a wide range of stakeholders. The differentiated nature of regulation should be proportional to the level of possible damage.
"AI needs to be regulated — neural networks are rapidly integrating into almost all areas of business and everyday life, and the rules of the game are needed here," Chernoutsyan said.
According to the expert, the real risks are inaccuracy in work, data leaks and "hallucinations" of AI. This is solved by engineering and organizational measures: data quality control, auditing and mandatory human supervision. The predominance of benefits over risks is a matter of proper management and AI settings.
On July 26, Polina Petrova, HR Director at Huntflow HR-tech company, told Izvestia that using AI to compile a resume could be harmful. According to her, thanks to AI, a clear and concise description is obtained. But at the same time, important details disappear from the resume — the context, the specifics of the tasks, the specific difficulties that the candidate faced.
All important news is on the Izvestia channel in the MAX messenger.
Переведено сервисом «Яндекс Переводчик»
