On Wednesday, Alexandr Wang, Meta’s chief artificial intelligence officer, announced a significant round of layoffs, impacting 600 employees within the company’s AI division. He stated that these cuts were a strategic move designed to accelerate Meta’s product development.
In an internal memo, Wang explained that a smaller team would streamline decision-making processes, leading to quicker innovation.
However, a less-publicized, yet equally impactful, series of layoffs occurred simultaneously. Meta dismissed over 100 individuals from its risk review organization. This department primarily comprised staff tasked with ensuring Meta’s compliance with Federal Trade Commission agreements and global privacy regulations.
Michel Protti, Meta’s chief privacy officer, later confirmed in a separate memo that the risk team would be downsized, with most manual review processes transitioning to automated systems.
Protti asserted that this shift from manual to automated reviews would lead to more accurate and reliable compliance outcomes. He reaffirmed Meta’s dedication to innovation while upholding its regulatory responsibilities.
Although the exact number of affected roles wasn’t disclosed by Protti, internal sources characterized these reductions as a severe blow to the teams responsible for evaluating privacy and integrity risks in Meta’s projects. The cuts notably affected the London office’s risk review team, with over 100 employees impacted across the company’s risk organization.
A Meta spokesperson released a statement clarifying that these organizational changes are routine and aim to reflect program maturity and boost innovation while maintaining high compliance standards.
These developments are part of a larger structural overhaul at Meta, driven by CEO Mark Zuckerberg’s three-year effort to re-energize the company and sharpen its competitive edge against emerging rivals, such as OpenAI, creators of the ChatGPT chatbot.
However, Meta executives reportedly grew impatient with the pace of product development. Ironically, one of the divisions contributing to this slower pace, by its very nature, was the risk organization itself.
Back in 2019, the Federal Trade Commission mandated that Meta, then operating as Facebook, implement new roles and practices to enhance transparency and accountability in handling user data. This was coupled with a record $5 billion fine against Facebook for misleading users about their control over personal data privacy.
The risk organization’s core duty was to oversee and audit all new products for potential privacy threats or any modifications that might breach the FTC consent order established in 2019.
In 2020, Protti had previously stated that these changes would usher in ‘a new level of accountability,’ emphasizing that privacy was ‘everyone’s responsibility’ at Facebook.
Both current and former risk organization employees expressed skepticism regarding the effectiveness of automated systems in managing sensitive issues like user privacy. This concern comes after Meta has spent nearly a decade under close surveillance by the FTC and Justice Department in the US, alongside significant scrutiny from European regulators.
For the past year, Meta had been gradually integrating automation into its risk auditing. This involved categorizing potential issues: ‘low risk’ updates underwent automated review followed by human audit, while ‘high or novel risk’ issues received immediate human auditor attention.
Last August, Meta reorganized its artificial intelligence division into four distinct units: FAIR, dedicated to research; TBD Labs, focused on advanced AI systems; a division for new product development; and an infrastructure group handling data centers and AI hardware.
Beyond the risk organization, Wednesday’s layoffs also impacted long-standing members of Meta’s FAIR team and individuals involved in previous iterations of Meta’s open-source AI models, known as Llama. Notably, Yuandong Tian, FAIR’s research director of eight years, was among those laid off.
Curiously, one division remained untouched by the layoffs: TBD Labs. This group primarily consists of new, highly compensated recruits dedicated to cutting-edge AI research and is headed by Alexandr Wang.