OpenAI has announced the establishment of a board committee to evaluate the safety and security of its artificial intelligence models, a significant governance change implemented just weeks after the resignation of its top safety executive and the subsequent dissolution of the internal safety team. The move has been met with varied reactions within the AI community and comes in the wake of criticism from former board members about OpenAI’s governance practices.
The newly-formed committee, which comprises directors and executives, has a 90-day mandate to assess the safeguards in place for OpenAI’s AI technology. The committee’s findings will be reviewed by the full board, and OpenAI plans to publicly share an update on the recommendations that get adopted.
This decision follows public criticism by Helen Toner and Tasha McCauley, former members of OpenAI’s board, who argued that AI firms should not govern themselves without adequate oversight. Their viewpoints, published in an opinion piece in The Economist, reflect growing concerns about the potential risks posed by advanced AI systems without robust governance mechanisms.
In a related move, OpenAI has transferred its ‘superalignment’ work, which focuses on ensuring that AI systems are aligned with human values, under the direct supervision of co-founder Greg Brockman. This reassignment highlights the company’s ongoing efforts to refine its approach to AI safety and alignment amid evolving industry dynamics and regulatory scrutiny.
For more details, please refer to the original article.