OpenAI’s Sam Altman Promises Ethics in Pentagon Deal — But Can He Deliver?

Sam Altman’s announcement of a Pentagon contract for OpenAI came with bold promises: no mass surveillance, no autonomous weapons, human control over lethal force. These are the very principles that got Anthropic banned from government contracts. The question the industry is now asking is whether OpenAI’s promises will prove more durable than Anthropic’s.
Anthropic spent months trying to negotiate a government AI deal that respected its core ethical guidelines, only to be publicly condemned by President Trump and effectively blacklisted from federal contracts. The company’s crime, in the administration’s eyes, was refusing to offer the Pentagon unfettered access to its AI regardless of how it would be used.
Altman stepped into the opening with characteristic speed, announcing both a Pentagon deal and a $110 billion funding round on the same night. His internal memo to employees struck a conciliatory tone, acknowledging the industry-wide implications of the Anthropic situation while stressing that OpenAI’s own ethical limits remained intact.
The skeptics, however, are numerous. Nearly 500 employees from across OpenAI and Google signed a solidarity letter with Anthropic, warning that the Pentagon was using financial incentives and political pressure to divide the industry and extract compliance from companies one by one.
Whether OpenAI’s deal truly protects against the uses that Anthropic refused will only become clear over time. The agreement as described by Altman sounds reassuringly similar to what Anthropic demanded — but Anthropic was punished for making the same demands, which raises uncomfortable questions about what, if anything, has actually changed.