San Francisco, September 16
Meta CEO Mark Zuckerberg has pushed back against growing calls to slow or pause artificial intelligence development, arguing that market incentives and legal liability can provide important safeguards against the technology’s risks.
“People won’t want to use agents that are misaligned with them and that don’t do what they ask, so labs have a strong natural incentive to make their models more aligned,” Zuckerberg wrote on X on Tuesday.
In AI development, “alignment” refers to efforts to ensure that an AI system’s goals and behaviour correspond with what humans intend.
The debate intensified in July after OpenAI, the maker of ChatGPT, disclosed that hundreds of AI agents had escaped their testing environment, connected to the internet and infiltrated a website during testing.
Zuckerberg argued that building trustworthy and reliable systems is becoming increasingly important for AI companies seeking to compete.
“Any lab that doesn’t focus on alignment will fall behind,” he said.
The debate has also expanded into questions over whether AI companies can be trusted to regulate themselves. OpenAI, Anthropic and Google are among companies exploring the creation of a joint standards body.
At the same time, prominent technology figures, including Anthropic CEO Dario Amodei, have called for AI development to be slowed.
Zuckerberg points to Meta’s AI rollout
Zuckerberg said legal liability would ultimately encourage AI companies to take responsibility for the risks associated with their products.
He cited Meta’s decision to delay the rollout of its Muse AI system, which includes personalised AI agents, by several months while the company worked on safety and security measures.
“We didn’t call for everyone else to do this before we would,” Zuckerberg wrote about the decision to delay the Muse AI release. “We just did it as part of our day-to-day work because it was clearly the right thing for people and for us.”
Meta has itself faced a series of liability lawsuits in recent months unrelated to AI. In August, the company agreed to pay billions of dollars to settle claims that Facebook and Instagram were designed to be addictive to children.
Zuckerberg also backed the use of independent evaluators to assess AI systems.
He said Meta Superintelligence Labs, the company’s AI division, already uses external reviewers and called for a “larger and more diverse ecosystem of evaluators” across the industry.
His comments came amid a broader debate over how AI companies should manage the risks posed by increasingly capable systems and whether regulation, industry standards or independent oversight should play the central role.
Zuckerberg’s post was also endorsed by several Silicon Valley figures close to the White House who have pushed back against warnings about the potential dangers of AI and calls for a slowdown in development.
