Meta: Internal Chaos in Its AI Strategy Sparks Criticism and Doubts

What happened?

According to a Wired report based on internal sources and reviewed meetings, Meta's new artificial intelligence unit —led by Mark Zuckerberg— is mired in chaos. In an employee meeting, a worker interrupted Zuckerberg to express frustration, reflecting widespread discontent. The company's AI strategy is perceived as disorganized, with constant priority shifts, duplication of efforts across teams, and a lack of clarity on long-term goals. This episode is not isolated: sources indicate that the lack of direction has caused key projects to stall or restart, while teams compete for limited resources. The situation echoes the internal turbulence Google experienced in 2018 with its AI strategy, before Sundar Pichai centralized efforts under Google AI. However, Meta lacks a similar unifying figure, and Zuckerberg's direct intervention, often erratic, has generated more confusion than clarity.

Why is it important?

Meta competes with giants like Google, OpenAI, and Microsoft in the generative AI field, a market that according to Grand View Research will reach $1.3 trillion by 2032. If its internal strategy is a disaster, the company risks falling behind, losing key talent, and diluting its multi-billion dollar AI investment. According to LinkedIn data, turnover in Meta's AI teams has increased 20% in the past year, and several high-profile scientists have migrated to OpenAI, Anthropic, and Google DeepMind. Additionally, investor and developer confidence could erode: Meta's shares fell 3% after the Wired report was published, though they partially recovered later. In the advertising market, which represents 98% of Meta's revenue, AI is crucial for ad targeting and optimization; a delay in tools like Meta AI or creator assistants could give an edge to competitors like TikTok and Snapchat, which are also heavily investing in generative AI.

Expected consequences

• Brain drain: Frustrated engineers and data scientists may migrate to more stable competitors. In recent months, Meta has lost several key researchers, such as the co-author of the Llama 2 model, who joined Anthropic. The lack of a clear vision discourages top talent, who seek projects with defined impact.
• Product delays: The lack of direction can delay the launch of new AI tools, such as advanced chatbots or creator assistants. For example, the Meta AI assistant, launched in September 2023, has received mixed reviews and its evolution has stalled. Additionally, the next version of Llama, expected for mid-2024, could be delayed, giving an advantage to open-source models like Mistral or Falcon.
• Regulatory criticism: A chaotic strategy could lead to hasty decisions that violate safety or ethical standards, attracting regulator attention. Meta has already been fined $1.3 billion by the EU for privacy violations, and an AI incident could worsen its situation. The European Commission is developing the AI Act, which will impose strict requirements on generative models; Meta may struggle to comply if its internal process is disorganized.

What readers should know

Meta is no stranger to drastic shifts in direction; recall its pivot to the metaverse in 2021, which cost it over $36 billion in operating losses through 2023. However, in AI, competition is fierce and the margin for error is minimal. The company urgently needs a clear roadmap, effective internal communication, and leadership that unifies teams. Unlike its bet on the metaverse, where Meta had a first-mover advantage, in AI it arrives late compared to OpenAI and Google. The success of its open-source strategy with Llama has been a bright spot, but internal chaos could jeopardize that advantage. Users and developers should monitor Meta's next moves, especially regarding open-source models like Llama, which could be affected by this internal chaos. If Meta fails to stabilize, it could miss the opportunity to be a relevant player in the next wave of AI, relegated to a secondary role in a market poised to transform the global economy.