Sam Altman Admits Mistake in Pentagon Deal: ‘It Looked Opportunistic and Sloppy’ - OpenAI Amends Contract After Backlash

calendar Mar 3, 2026 · 4 min read · openai pentagon sam-altman anthropic surveillance ai-ethics military  ·
Share on: linkedin copy

Sam Altman Admits Mistake in Pentagon Deal: “It Looked Opportunistic and Sloppy”

OpenAI CEO Sam Altman publicly admitted to “getting it wrong” by rushing the Pentagon agreement, promising to add stricter clauses to the contract following a wave of criticism and protests.

What Happened

In a series of posts on X on Monday (03/02), Altman shared an internal memo acknowledging that the company shouldn’t have rushed the announcement of the Department of War deal:

“One thing I think I did wrong: we shouldn’t have rushed to get this out on Friday. The issues are super complex and demand clear communication. We were genuinely trying to de-escalate and avoid a worse outcome, but I think it just looked opportunistic and sloppy.”

The admission came after a week of intense backlash against OpenAI, including:

  • Protests in San Francisco and London
  • A boycott organized by the QuitGPT group
  • Mass cancellations of ChatGPT subscriptions
  • Approximately 500 OpenAI and Google employees signing an open letter supporting Anthropic

Clauses Added to Contract

In response to the criticism, OpenAI announced it is amending the agreement with the Pentagon to add more explicit language about domestic surveillance:

New Contractual Language

“Consistent with applicable laws, including the Fourth Amendment to the United States Constitution, National Security Act of 1947, and FISA Act of 1978, the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals.”

Prohibition of Use by Intelligence Agencies

Altman confirmed that the Pentagon also affirmed that OpenAI’s services will not be used by War Department intelligence agencies (for example, the NSA), and that any service to those agencies would require a follow-on modification to the contract.

Protection Against Commercial Data Surveillance

The amendment also specifies that the prohibition includes “deliberate tracking, surveillance, or monitoring of U.S. persons or nationals, including through the procurement or use of commercially acquired personal or identifiable information.”

Context: The OpenAI vs Anthropic vs Pentagon Saga

This retreat is a significant development in the dispute that began on Friday (02/27):

  1. Anthropic rejects Pentagon terms - The company refused to sign a $200 million contract without explicit guarantees against domestic mass surveillance and autonomous weapons
  2. Trump bans Anthropic - The president ordered federal agencies to stop using Anthropic’s technology
  3. Pentagon designates Anthropic as “supply chain risk” - A label typically reserved for foreign adversaries
  4. OpenAI announces deal hours later - OpenAI announced its agreement on the same Friday, generating criticism about “opportunism”
  5. Public backlash - Protests, subscription cancellations, and a letter from 500 employees
  6. Altman admits mistake - The public admission and contract amendments represent a significant retreat

Pentagon’s Reaction

In a statement, the Pentagon stated that:

“We will always come to the table for reasonable discussion as we did with OpenAI. Anthropic didn’t want to do that, because they have their own personal vendettas.”

What This Means

Altman’s admission and the contract amendments represent an important turning point:

Partial Victory for Anthropic’s Ethical Stance

Anthropic insisted on stricter guarantees about domestic surveillance — now OpenAI is being forced to add similar clauses under public pressure.

Consumer and Employee Pressure Worked

The boycott, protests, and letter from 500 employees demonstrate that consumers and workers have power to influence major AI company decisions.

Precedent for Transparency

Altman’s admission of error is rare in the tech industry and may set a precedent for greater transparency and accountability in controversial AI decisions.

Market Impact

According to reports, there was a 295% increase in ChatGPT uninstalls following the initial Pentagon deal announcement, with many users migrating to Anthropic’s Claude.

Altman acknowledged the impact in his weekend conversations, reiterating that Anthropic should not be designated as a “supply chain risk” and hoping the Pentagon offers the company the same terms now agreed upon with OpenAI.

Sources