Zuckerberg Backs Wider Access to Superintelligence
Meta’s chief says advanced AI should be widely available, with independent directors overseeing safety standards for model releases.
Topics
News
- Temasek, Seraphim Back Pixxel in $100 Million Round
- UN Rights Chief Warns Advanced AI Could Threaten Humanity
- Tata-owned JLR to Cut About 4,000 Bobs Amid Tariff, China Pressure
- Anthropic Walks Away From $6 Billion Decart Deal
- Hyundai India Targets 20% Women in Executive Workforce by 2030
- Security Breach Drains $320 Million in Bitcoin From Liquid Network
Meta Chief Executive Mark Zuckerberg said the company would seek to make superintelligent AI widely available rather than allow governments, large companies or a handful of developers to control it.
In a policy statement published on Monday, 10 August, Zuckerberg said Meta would focus on “personal superintelligence,” including AI agents that could assist with health, careers, finances, relationships and household tasks.
Meta plans to offer free versions of the tools to billions of people, while charging users who want more computing capacity. Zuckerberg also said the company would build a private mode in which even Meta could not see or grant access to users’ information.
Superintelligent AI remains a prospective technology, and Zuckerberg did not set a timetable for delivering the agents described in his statement.
Meta will also restart some open-source model releases following the establishment of Meta Superintelligence Labs.
“Meta continues to be strongly supportive of open source, including open source AI models,” Zuckerberg wrote. “We will resume releasing some open source models soon.”
The announcement comes as AI companies and policymakers debate whether increasingly capable systems should face tighter restrictions because of risks involving cybersecurity, biological research and autonomous agents.
Zuckerberg argued that restricting advanced AI to a small number of institutions could create a different kind of risk by concentrating power. He said wider access could allow individuals, companies and governments to act as checks on one another.
Meta is also changing how it oversees releases of its most advanced models. Zuckerberg said the company’s independent board directors will be given authority to approve safety criteria and assess whether models meet those standards before release.
“I do not think it is in my, Meta’s, or the world’s best interests for me or anyone else to be a sole decision-maker on how superintelligence is deployed,” he said.
Zuckerberg also proposed closer cooperation between frontier AI companies and the US government, including giving agencies access to intermediate model-training checkpoints so they can assess emerging capabilities and strengthen critical systems.
At the same time, he warned against regulation that significantly delays US. AI development, arguing that competition with China is moving quickly.
On employment, Zuckerberg rejected predictions that superintelligence would inevitably eliminate large numbers of jobs. He said Meta would emphasize AI as a tool for invention and individual productivity, predicting that smaller teams could eventually create and operate businesses that currently require much larger workforces.
