President Donald Trump has been very vocal about how much he thinks he deserves the Nobel Peace Prize. But so far, he has been shut out.<\/p>\n
Despite launching a war against Iran early this year, however, concerns about AI\u2019s destructive potential could provide an opening for him to win the award. <\/p>\n
In an exclusive interview with\u00a0Fortune<\/a>\u00a0on Friday, OpenAI CEO Sam Altman was asked about the AI industry slowing down in the name of safety and if that extends to the U.S.-China competition.<\/p>\n \u201cI think Presidents Trump and Xi would get the Nobel Peace Prize together if they could agree on something that should be easy to agree to, and it would be wonderful,\u201d he told Fortune\u00a0Editor-in-Chief Alyson Shontell.<\/p>\n The two leaders ought to concede that neither side should take a certain level of risk in the AI development process, Altman added.<\/p>\n \u201cEven if just the U.S. and China could agree on some shared standards and testing for development of this technology, I think that\u2019d be a wonderful accomplishment that the two of them can deliver,\u201d he explained.<\/p>\n When asked if a ban on recursive self-improvement\u2014meaning an AI that can continuously upgrade itself and add more capabilities\u2014would be enough, Altman replied that it probably would not be sufficient.<\/p>\n Still, he maintained that a notional U.S.-China pact on AI isn\u2019t hard to accomplish and could even be laid out in a one-page document.<\/p>\n The key is having guardrails in place to mitigate the risk of losing control over an AI while also not letting safety measures leave one side at a disadvantage.<\/p>\n \u201cIn the extreme case, the U.S. and China are both worrying that if one country gets superintelligence first, that\u2019s a big imbalance of power in that relationship,\u201d Altman explained. \u201cOn the other hand, neither of us should be taking loss-of-control risk in the race to beat the other country.\u201d<\/p>\n In his view, the optimal terms for an AI deal signed by the world\u2019s two superpowers could look something like this:<\/p>\n \u201cHere are the rules that we need on development and testing and monitoring alignment standards before anyone proceeds. And then here is the oversight that both countries will have, or some international body will have, to ensure that we are avoiding one country running away in the concentration-of-power element or not following the rules in the first place.\u201d<\/p>\n Such talk about a U.S.-China treaty comes as the stakes in the AI race have reached existential levels, on par with the nuclear arms race during the Cold War.<\/p>\n While earlier concerns about AI mostly focused on jobs that are at risk, students cheating on assignments, and when the hyperscaler investment boom will go bust, anxiety has recently turned to whether humans will survive AI.<\/p>\n On Tuesday, a researcher at Anthropic publicly announced his resignation, and accused both Anthropic and OpenAI (where he also previously worked) of acting irresponsibly in developing ever more capable AI systems.\u00a0\u00a0<\/p>\n \u201cThey are racing straight to self-improving superintelligence and gambling with our lives,\u201d wrote Jacob Coxon, the researcher, in a post on X.\u00a0Anthropic\u2019s head of alignment commented\u00a0<\/a>on the post, saying Coxon was correct in his assertion that many Anthropic and OpenAI researchers believe that increasingly powerful AI could potentially wipe out humanity. <\/p>\n Evan Hubinger, Anthropic\u2019s \u201calignment science lead,\u201d wrote in response to Coxon\u2019s resignation post that \u201cwe really do earnestly believe AI could kill all humans,\u201d putting the odds of that happening within the next decade at more than 10%.<\/p>\n On Saturday, Anthropic CEO Dario Amodei\u00a0announced\u00a0that his company is\u00a0committing to a new safety measure<\/a>\u2014giving independent evaluators permanent, employee-level access inside the company\u2014as part of a broader plan he says is needed to slow the pace of AI development.<\/p>\n For his part, Altman also suggested OpenAI and other leading AI companies\u00a0may be close to announcing a pact to slow AI development<\/a>\u00a0and collectively address the rapidly increasing safety risks. While the AI labs \u201cfeel the gravity of this moment,\u201d he told Fortune on Friday that the geopolitics between the U.S. and China represent a different dynamic.<\/p>\n Paul Tudor Jones, founder of Just Capital and of Tudor Investment Corp., similarly said the AI threat \u00a0demands coordinated action by the two countries, flagging recursive self-improvement in particular as a key threshold.<\/p>\n