Activists Demand OpenAI Withdraw Pentagon Contract

A petition on OneGreenPlanet is calling on Sam Altman to withdraw OpenAI from a contract with the Department of Defense. The petition frames the move as opportunistic after Anthropic allegedly refused a similar agreement on ethical grounds and then faced federal restrictions. The core complaints are the potential for AI to power mass surveillance and autonomous weapons and the lack of enforceable safeguards once models are integrated into military systems.

The published piece does not spell out contract terms, but the operational risks for practitioners are clear: once models or APIs are embedded into defense systems, downstream control becomes limited and standard mitigation practices can be ineffective. Key technical risks include:

• •Dual-use deployment: general-purpose models can be repurposed for targeting, surveillance, or automated lethal decision-making.
• •Loss of oversight: API access and model updates complicate logging, provenance, and auditability inside classified systems.
• •Data and model leakage: integration with military pipelines increases chances of sensitive training or operational data being exposed.

This episode sits at the intersection of procurement, corporate ethics, and AI governance. Tech firms face conflicting pressures: lucrative government contracts versus developer, customer, and public expectations about safety and nonparticipation in offensive uses. The Anthropic refusal and the petition spotlight how procurement actions can reshape norms about acceptable use and trigger policy responses, including agency-level bans or congressional oversight.

Monitor OpenAI's public response, any disclosure of contract scope or usage constraints, activist traction, and whether regulators or procurement offices impose stricter requirements for verifiable, enforceable safeguards.