OpenAI wants the US government to take the lead on writing global rules for AI systems that could someday upgrade themselves.
In a proposal published Monday, September 21, OpenAI called on the United States to lead an international effort, working with other countries, to set standards around recursive self-improvement, the process by which an AI system enhances its own capabilities with little human oversight. The company was careful to draw a line under its ambitions. "Fully autonomous RSI is not happening today, and we should not pursue it unless and until it can be done safely," OpenAI wrote. Rather than describing self-improving systems as imminent, the proposal frames this as getting ahead of a technology OpenAI itself may eventually build.
A company asking Washington to lead global rules for a technology it is racing toward is a familiar pattern. OpenAI has spent years positioning itself as the industry's safety-conscious voice while continuing to push capability forward, and putting the US in the driver's seat on standards conveniently favors the country where OpenAI and most of its well-funded rivals are headquartered.
Until a genuinely self-improving system exists, this reads like a company drafting the exam questions before anyone has to sit the test.