Internal Slack messages at OpenAI show that employee pushback and worries about reputational damage led Greg Brockman, OpenAI’s president, to withdraw the second half of a publicly announced $50 million pledge to the super PAC Leading the Future (LTF). The debate centered on LTF’s tactics, questions about its independence, and staff concerns that aggressive political activity could drive researchers away and harm the company’s standing on AI safety.
What happened and when
- On June 1, 2025, Jason Kwon, OpenAI’s chief strategy officer, shared a preview of a blog post in an internal Slack message asserting that OpenAI had no formal connection to Leading the Future and that Greg Brockman’s support had been "in a personal capacity, not on behalf of the company."
- Employees reacted critically, warning that the company’s political activity risked pushing talent to competitors and undermining OpenAI’s position in the race toward recursively self-improving AI.
- Ten days later, on June 11, 2025, Brockman announced in Slack that he would not follow through on the second $25 million installment of his $50 million public commitment, writing that he felt "terrible" about how LTF reflected on the company.
Key threads in the internal debate
- Some staff doubted Kwon’s framing of LTF as independent. One worker noted that Brockman’s contribution could fund an X account they called an "LTF sockpuppet," which posted inflammatory material that alarmed colleagues.
- A source familiar with the situation told Semafor that employee opposition was definitely part of Brockman’s decision, but that he also had separate concerns after LTF targeted Alex Bores, a New York Assemblymember and AI-safety advocate running for Congress; Brockman had expected the PAC to focus more on allied candidates.
- Kwon later acknowledged reputational hits in Slack and said he wished the company had drawn clearer public lines sooner about the political tactics it does not condone.
Leading the Future and related organizations
- Leading the Future launched in August 2025 and announced more than $100 million in commitments, including $50 million from Brockman and his wife, Anna.
- Chris Lehane, OpenAI’s head of global affairs, helped set up the PAC, using a model he previously deployed in the crypto sector.
- LTF has made donations to Build American AI (BAAI), a sister lobbying group that does not disclose donors. OpenAI told journalists it had not funded those groups.
Political effects and contradictions
- In practice, LTF has largely donated to frontrunners in congressional races. Some candidates it supported have backed legislation proposed by Sen. Bernie Sanders and Rep. Greg Casar that would pause frontier AI development and criminalize attempts to build superintelligence — a stance that conflicts with OpenAI’s stated aims.
- Separately, Brockman had given another political donation to MAGA Inc.; one source suggested that donation could have been a larger headache, but employees’ objections centered on LTF’s conduct.
Inside impact and broader implications
- The Slack messages were shared with Semafor on condition that direct quotes not be reproduced verbatim, to avoid chilling employees’ willingness to air concerns internally in the future.
- The episode illustrates that many OpenAI developers, engineers and researchers are skeptical of unconstrained AI development and dislike being associated with aggressive political tactics.
- After internal and external pressure, OpenAI has started disclosing examples of concerning model behavior; the company also said it would be more explicit about political tactics it does not support.
Conclusion
Employee activism and concerns about the company’s reputation contributed to Brockman’s decision to cancel the second $25 million pledge to Leading the Future. The episode offers a rare view into tensions at the companies shaping AI policy: staff retention, safety worries and public accountability can directly influence corporate political commitments. OpenAI declined to comment.
