In 2017, OpenAI co-founder Greg Brockman proposed that the company could enrich itself by playing world powers - including China and Russia - against one another, perhaps by starting a bidding war among them. The idea emerged during internal discussions about how to fund the massive compute requirements of artificial general intelligence, and it horrified the company's first policy adviser.
Page Hedley, a former public-interest lawyer hired by Dario Amodei to advise on ethics, had presented executives with a plan for averting a catastrophic AI arms race - possibly through a NATO-like coalition of labs, "Hvylya" reports, citing a New Yorker investigation. Brockman's response, according to Hedley and contemporaneous records, was to ask how OpenAI could raise more money and win. His counterproposal: sell AI capabilities to the highest bidder among nation-states. "The premise, which they didn't dispute, was 'We're talking about potentially the most destructive technology ever invented - what if we sold it to Putin?' " Hedley recalled.
The proposal, known internally as the "countries plan," did not die in a brainstorming session. According to several people involved and contemporaneous documents, OpenAI executives grew more excited about it over time. Jack Clark, the company's policy director at the time, said Brockman's goal was to "set up, basically, a prisoner's dilemma, where all of the nations need to give us funding, and that implicitly makes not giving us funding kind of dangerous." Executives discussed the approach with at least one potential donor.
A junior researcher recalled thinking at a company meeting where the plan was detailed: "This is completely fucking insane." Several employees talked about quitting. The plan was ultimately abandoned - not because executives recognized it as reckless, but because Altman feared losing staff. "I feel like that was always something that had more weight in Sam's calculations than 'This is not a good plan because it might cause a war between great powers,' " Hedley said.
OpenAI disputed the characterization. "Ideas were batted around at a high level about what potential frameworks might look like to encourage cooperation between nations - something akin to an International Space Station for AI," a representative said. "Attempting to characterize it as anything more than that is utterly ridiculous." Brockman maintains he never seriously entertained auctioning AI models to governments.
Also read: how China's economic espionage dwarfs anything pre-WWI Germany achieved.