Zuckerberg of Meta (META.US) responds to AI "slowdown" debate: Continue advancing model capabilities and strengthening security protection

According to the Zhitong Finance APP, Meta Platforms (META.US) CEO Mark Zuckerberg and the company's AI chief Alexander Wang have spoken out on the debate over the pace of AI development, advocating that AI labs should continue to advance model capabilities while simultaneously building stronger safeguards around increasingly powerful models.

The pair made these remarks on the X platform. The industry is currently engaged in a broad debate about whether the development of AI capabilities should be slowed until more progress is made in AI safety. AI safety refers to ensuring that AI systems reliably follow human intentions and values.

As of Tuesday, Meta’s stock price had risen for three consecutive trading days, up a total of 4.2%, with slight gains in after-hours trading. On Stocktwits, as of the last check, the retail sentiment towards the stock was "extremely bearish."

Zuckerberg: Labs Are Motivated to Build Safer AI

Zuckerberg stated that AI companies already have strong commercial and legal incentives to prioritize model safety and alignment with users. "People won’t want to use agents that are not aligned with them or don’t do what they ask," Zuckerberg wrote. "Any lab that doesn’t focus on alignment will fall behind."

He believes alignment itself is becoming an important competitive capability, not just a constraint on AI development. "My view is that trust and alignment are rapidly becoming the most important abilities distinguishing agents and models," he said.

Zuckerberg also mentioned that Meta had decided to delay the release of the Muse model by several months in order to address safety and security issues.

Meta Supports External Oversight

Both Zuckerberg and Alexander Wang have called for the introduction of more independent evaluators and external oversight as AI models become increasingly powerful.

"Bringing in independent evaluators and consultants is industry best practice," Zuckerberg said, and he believes a larger and more diverse community of evaluators will improve AI safety efforts.

Alexander Wang similarly stated that AI labs should maintain "strong governance frameworks" throughout the entire training and deployment process, including external evaluators and independent oversight of safety standards for model release.

He added: "Every lab must operate under the protections of democratic institutions... This means that if a model causes harm, the lab will face significant legal liability."

Meta Defends Computing Power Allocation Strategy

A key part of the two executives’ argument is that Meta has decided to allocate most of its computing power to serve users, rather than accelerating recursive self-improvement (RSI).

"Investing the vast majority of our computing resources in serving humanity, rather than chasing recursive self-improvement, is one of the best ways to ensure we develop this technology safely," Zuckerberg said.

The concern is that RSI could allow AI systems to rapidly enhance their own capabilities, potentially beyond human supervision, and increase the risk of powerful models running out of control.

Alexander Wang also described the pursuit of RSI as "one of the most dangerous paths that could lead to powerful models getting out of control."

AI Development Pace Debate Splits the Tech Industry

The debate over the pace of AI development is roughly divided: Anthropic CEO Dario Amodei has called for slowing down frontier AI development to let safety mechanisms catch up; OpenAI CEO Sam Altman, Microsoft CEO Satya Nadella, and SpaceX’s Elon Musk support proceeding more cautiously and with a focus on safety—without halting development.

On the other hand, U.S. President Donald Trump opposes slowing America’s AI race, citing concerns about China taking the lead; former Trump AI advisor David Sacks believes companies should manage risks themselves rather than seek government restrictions, and has questioned Anthropic’s safety warnings.

Investor Michael Burry has questioned whether calls to slow development are partly self-serving, as this might protect existing AI companies from competitive shocks.