OpenAI has confirmed that it held discussions with rival AI companies Anthropic and Google DeepMind on AI safety, as concerns grow over the risks posed by increasingly capable artificial intelligence models.
Chris Lehane, OpenAI’s global policy chief, told CNBC in Washington DC on 15 September that the discussions had taken place over several weeks. He said OpenAI did not believe the companies needed an antitrust waiver to coordinate on AI safety, arguing that collaboration was preferable to pursuing separate approaches.
Lehane also said OpenAI would support bipartisan legislation aimed at mitigating catastrophic AI risks.
Anthropic and Google DeepMind did not respond to CNBC’s requests for comment, making OpenAI the only company to confirm the discussions.
The talks come amid growing calls from some AI industry leaders for a slower pace of frontier AI development.
Anthropic CEO Dario Amodei argued in an essay published on 12 September that frontier laboratories should deliberately slow capability gains. He warned that increasingly capable AI agents could potentially take over the internet through a persistent botnet within six to 12 months.
Amodei proposed placing independent evaluators inside AI laboratories, establishing common safety standards among developers in democratic countries and pursuing international regulatory coordination.
OpenAI CEO Sam Altman backed the proposal to use independent evaluators with employee-like access, describing the idea as “great” and saying OpenAI would adopt the approach.
Altman later clarified that pacing frontier AI development did not mean stopping development. He also ruled out an OpenAI public listing in 2026, telling Fortune that safety concerns made it an ill-advised time for an initial public offering.
Trump rejects AI slowdown
US President Donald Trump has rejected calls for a slowdown in frontier AI development, arguing that “very negative forces” were exaggerating the risks and that the US needed to maintain its AI lead over China.
China has also dismissed the proposal. Its foreign ministry called for an “open, inclusive and benevolent approach” to AI, while state-run Global Times described Amodei’s plan as a “Cold War playbook” aimed at containing China’s ambitions.
OpenAI said its discussions with Anthropic and Google DeepMind began before Amodei published his essay and before a former Anthropic researcher publicly resigned over the approach taken by leading AI laboratories.
The discussions also follow a July proposal from Google DeepMind chair Demis Hassabis for a US-led frontier AI standards body modelled on the Financial Industry Regulatory Authority.
Hassabis proposed a public-private organisation funded by industry and staffed by independent technical experts to evaluate advanced AI models before their release.
