OpenAI and Anthropic Back Global AI Pause Watchdog
Key insights
- OpenAI's memo, authored by CEO Sam Altman and chief scientist Jakub Pachocki, calls for a global body able to pause frontier AI development.
- OpenAI projects that by March 2028, AI systems could conduct a significant fraction of its own research alongside human researchers.
- Anthropic, which recently surpassed OpenAI as the world's most valuable startup, said a global AI slowdown would 'likely be a good thing.'
Why this matters
When competing frontier labs, both heading toward 2026 IPOs, simultaneously call for external limits on their own development pace, it hands international regulators a direct foothold for formal oversight mechanisms. OpenAI's March 2028 self-projection, in which AI could be conducting a significant fraction of frontier research, gives governments a concrete timeline to work backward from when drafting policy. The fact that Anthropic, now the world's most valuable startup, explicitly endorses a potential global pause reframes the standard industry argument that commercial success and safety goals are inherently incompatible.
Summary
OpenAI CEO Sam Altman and chief scientist Jakub Pachocki published a memo calling for an international body to oversee AI, with explicit authority to slow frontier progress when safety falls behind. Anthropic made the same argument one week earlier.
Both labs cite recursive self-improvement as the key risk, where AI systems train increasingly capable versions of themselves. OpenAI projects that by March 2028, AI could conduct a significant fraction of its own research alongside human researchers.
Essentially: OpenAI and Anthropic, both heading toward stock market debuts in 2026, have now publicly aligned on external AI governance.
- OpenAI calls for a body able to coordinate a global development pause when safety lags behind progress.
- Anthropic said a global slowdown would 'likely be a good thing.'
- Anthropic recently surpassed OpenAI as the world's most valuable startup.
Two direct rivals publicly calling for constraints on their own development pace is a signal regulators have been slow to respond to.
Potential risks and opportunities
Risks
- Both companies face regulatory capture accusations if they help design an oversight body while being its primary regulated entities, complicating their 2026 IPO disclosures
- OpenAI's March 2028 timeline for AI-led research, if reached before any governance framework exists, could trigger emergency regulatory action that disrupts product roadmaps
- Anthropic, now the world's most valuable startup, faces post-IPO shareholder pressure to walk back voluntary slowdown commitments if global coordination stalls
Opportunities
- International standards bodies such as ISO and NIST gain direct leverage with both labs, who have now publicly endorsed the concept of formal international AI oversight
- AI safety and alignment research organizations are positioned to supply the technical expertise a new international AI oversight body would require
- Policy and regulatory technology vendors could see accelerated government procurement cycles as multiple jurisdictions respond to the dual-lab governance call in 2026
What we don't know yet
- Whether Sam Altman's June 3, 2026 congressional meeting produced any specific legislative commitments toward the proposed international oversight body
- Which governments have indicated willingness to participate in or fund the kind of international AI oversight body both labs are now calling for
- Whether OpenAI's March 2028 projection that AI could conduct a 'significant fraction' of its own research is grounded in published benchmarks or undisclosed internal modeling
Originally reported by gizmodo.com
Read the original article →Original headline: OpenAI Joins Anthropic in Calling for International AI Watchdog With Authority to Coordinate a Global Pause on Frontier Development