OpenAI employees made their president’s political spending a talent problem

Semafor’s 2 October reporting reconstructs an internal dispute over Greg Brockman’s support for Leading the Future, a super PAC opposing AI regulation. The withdrawal happened in June; the new evidence is the staff discussion surrounding it.

After strategy chief Jason Kwon previewed a company statement on 1 June, an employee warned that the political association could drive researchers to Anthropic and damage OpenAI’s competitive position. Kwon subsequently acknowledged reputational damage. On 11 June, Brockman abandoned the remaining $25mn of a publicly announced $50mn commitment.

OpenAI’s public statement draws a distinction between employees’ personal political activity and company advocacy. It says no outside group speaks for OpenAI, denies directing Leading the Future, and supports regulation addressing testing, safety and accountability.

Semafor’s source says staff pressure contributed to Brockman’s decision, alongside his own concerns about the PAC attacking AI-safety advocate Alex Bores. OpenAI declined to comment to the publication. The messages therefore do not establish that employee objections were the sole cause.

Analysis: an executive’s formally personal donation can still affect a laboratory’s reputation and recruitment. The episode also complicates the idea that OpenAI’s workforce uniformly favours unrestricted acceleration: retaining scarce researchers can place a practical limit on its leaders’ political strategy.