OpenAI Tightens Pentagon Deal After Backlash, Bans Domestic Surveillance

OpenAI has agreed to amend its recently announced Pentagon deal after a backlash over the potential use of its AI in classified military operations. CEO Sam Altman said the changes add guardrails and explicitly prohibit using OpenAI systems to spy on Americans.

Under the revised terms, intelligence agencies such as the National Security Agency would need a follow-on modification to access OpenAI's technology. Altman also noted that the company wants clear safeguards and that these tweaks are designed to prevent domestic surveillance and to ensure any use aligns with strict congressional and policy guardrails.

Altman acknowledged the initial rollout was rushed, describing the situation as opportunistic and sloppy in hindsight. He said the company aimed to de-escalate tensions, but public reception suggested the move looked sudden and self-serving, prompting a pause to refine the agreement.

The Pentagon deal had sparked ongoing concerns about mass surveillance and the balance of power between government needs and private sector capabilities. Proponents argue guardrails are essential as AI is embedded into national security operations, while skeptics warn that governance must be robust and transparent.

OpenAI’s latest step may help rebuild trust among users and policymakers, even as observers monitor how strictly the new provisions will be enforced and whether additional clarifications will be required for future military deployments. The online reaction also reflected broader debate about AI in defense and the responsibilities of tech companies when engaging with government programs.

Some users reportedly reduced their use of ChatGPT following the news, illustrating the broader reputational impact of military partnerships on AI platforms.

#OpenAI #AIPolicy #NoMassSurveillance