Mark Zuckerberg argues that competition and user demand already push AI labs to prioritize safety, pushing back against calls for new regulation.
Meta CEO Mark Zuckerberg entered the discussion on AI extinction risks this week, contending that market forces already compel labs to focus on safety without needing fresh regulations.
Weeks earlier, the chief executive of Meta (META) had issued a warning about superintelligence. In his latest posts, he stated that the technology ought to empower individual users rather than a small group of companies.
In his essay from August, Zuckerberg argued against any single corporation or government controlling superintelligent systems, advocating for widespread distribution instead.
Last month I wrote about how we can build a positive and safe future for everyone: https://t.co/1KtONlqBs5
— Mark Zuckerberg (@finkd) September 15, 2026
Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens.
The…
Zuckerberg's posts came as policymakers in Washington contend with alerts from within the AI sector. This month, Anthropic researcher Jacob Coxon resigned, charging his own company and OpenAI with hastening self-improving systems while lacking proper protections.
The resignation intensified a campaign to prohibit superintelligence entirely. Senator Bernie Sanders and Representative Greg Casar are spearheading that initiative in Congress.
Separately, OpenAI safety researcher Marcus Williams estimated the probability of human extinction at 70% in three years. He added that the figure could decline if laboratories decelerate or if regulators step in.
According to Zuckerberg, alignment rather than restriction will distinguish sustainable AI agents from unsuccessful ones. He asserted that trust is emerging as the sector's most important asset.
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.
The Meta chief shared his views in a post on X.
Zuckerberg also highlighted Meta's safety record. The company postponed the launch of its Muse personal assistant from a planned April date to fix security issues. He maintained that Meta took that step independently, not awaiting sector-wide regulations.
He further noted that Meta Superintelligence Labs (MSL) already employs external evaluators to assess its work. He contended that competing labs could take up comparable independent assessments.
The founder of Meta added that channeling the majority of computing resources toward serving users rather than self-improvement ensures safe development.
Zuckerberg's position does not settle the fundamental dispute. Critics call for binding restrictions on self-improving AI, whereas Meta is counting on competition and self-interest.
How that equilibrium holds could influence Congress's approach to Zuckerberg's advocacy for rapid AI releases in the coming months.
Share to
Disclaimer: this article comes from third-party media and is provided for reference only. It does not constitute investment advice. Crypto and other financial products carry significant price volatility risk, so please make your own decisions carefully.
Japan's services PMI rose to 51.7 in October, but the composite PMI slipped to 52.1, a five-month low, as hiring slowed and cost pressures persisted.
September data showed Australia's goods surplus at about A$2.8 billion, a seven-month peak, with LNG and metals exports up and gold imports down sharply.
SNB Chair Martin Schlegel said October inflation beat central bank forecasts and signalled readiness to act, firming December rate-hike bets.
Minutes from the BoJ's September meeting showed policymakers agreed on more rate hikes, flagged a 50% oil surge risk and heard government calls for caution.