AI companies agree to voluntary safety commitments

Artificial intelligence companies agreed on Tuesday to a set of voluntary safety guidelines and reaffirmed their commitment to developing their AI technology in a way that incentivizes the security of their platforms and the public.

At the AI Seoul Summit – hosted by the South Korean and United Kingdom governments – 16 leading companies signed onto the non-binding Frontier AI Safety Commitments, which is aimed at mitigating the risks of AI while increasing accountability and transparency.

Among the companies that agreed to the commitments are OpenAI, Google, IBM, Amazon, Meta and xAI.

The companies agreed to assess the risks of the technology at every step of the development of a new AI system, focusing on reducing risks associated with possible misuse of the technology as well. They agreed to consider the results of internal and external evaluations “as appropriate” to determine the safety of the system. The external evaluations could be conducted by independent third parties, their home governments, or “other bodies their governments deem appropriate.”

The companies also agreed to set a pre-determined threshold for the maximum level of risk posed by a given AI system that they would be willing to tolerate. They agreed to state clearly how they would ensure they keep the risks under that threshold and what they would do if their AI system exceeded that threshold, including “processes to further develop and deploy their systems and models only if they assess that residual risks would stay below the thresholds.”

“In the extreme, organizations commit not to develop or deploy a model or system at all, if mitigations cannot be applied to keep risks below the thresholds,” read part of the non-binding agreement.

The two-day summit followed the first such gathering in November, and it featured several world leaders who similarly pledged to build proper guardrails to develop AI safely and to share resources to “forge a common understanding of AI safety and align their work on AI research.”

Tags AI UN

Copyright 2024 Nexstar Media Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed..

 

Main Area Top ↴

Testing Homepage Widget

 

Main Area Middle ↴
Main Area Bottom ↴

Most Popular

Load more

Video

See all Video