Meta restructures its AI divisions

Meta dissolved the team responsible for regulating and mitigating potential threats in AI development. The Information reports, citing a corporate memo.

Some employees from the dissolved unit joined the company’s Generative AI department, while others began work on the infrastructure-focused division.

Despite the internal reallocation, members of the ‘Responsible AI’ team remain committed to the principles of safe AI use, according to the documents.

The restructuring comes as Meta’s ‘Year of Efficiency’ draws to a close — as CEO Mark Zuckerberg called it in the the March Year of Efficiency report. He has repeatedly mentioned the need to redeploy resources and prune ‘several layers of management’.

“Since last year we reduced the workforce, one surprising result has been that many matters moved faster. Looking back, I underestimated the indirect costs of lower-priority projects,” Zuckerberg said.

In November 2022 Meta fired a team of machine-learning researchers in the AI infrastructure stack. Prior to that, the company announced a reduction of 11,000 employees.

Ensuring AI safety has become a priority area for leading players, given the intense scrutiny by regulators of the industry.

In May, the Biden administration held a meeting with the heads of Google, Microsoft, OpenAI and Anthropic to discuss AI-related risks. In the same month the former OpenAI CEO Sam Altman testified before the US Congress and urged the government to regulate the use and development of the technology.

Earlier in October, the Biden administration published a directive setting new standards for AI safety and protection. According to the document, irresponsible use of AI can lead to discrimination, bias, and other abuses.