Editorial illustration for OpenAI yields to Pentagon, bans bulk U.S. data; Amodei says law not yet
OpenAI Limits Pentagon Data Access, Blocks Bulk Processing
OpenAI yields to Pentagon, bans bulk U.S. data; Amodei says law not yet
The Pentagon wants OpenAI’s power, but not the moral headache that comes with it. So the company has drawn a line: no bulk collection of Americans’ data, no open-ended surveillance, no generalized harvesting of the public. Read the fine print, though, and that line starts to look like a suggestion.
Anthropic’s Dario Amodei puts it plainly: the law hasn’t caught up to what AI can already do at scale. OpenAI’s own contract hedges with words like “unconstrained” and “generalized”, language that grants flexibility, not a firewall. Sarah Shoker, a former OpenAI researcher, calls it designed for optionality.
The result? Under current rules, the Pentagon could legally use the same system to search foreign intelligence databases for information on Americans, en masse. The ban is real, but so is the loophole.
In practical terms, this means the system cannot be used to collect or analyze Americans' data in a bulk, open-ended, or generalized way." Anthropic's Amodei has publicly said that the law had not yet caught up with AI's ability to conduct surveillance on a massive scale. And Altman takes pains in his statement to say that OpenAI's contract "reflects [its red lines] in law and policy," meaning that it's simply abiding by existing laws and existing Pentagon policies, the latter of which can change at any time. (OpenAI attempts to address the latter issue in a Q&A, where it says the contract "explicitly references the surveillance and autonomous weapons laws and policies as they exist today, so that even if those laws or policies change in the future, use of our systems must still remain aligned with the current standards reflected in the agreement.") Sarah Shoker, a senior research scholar at the University of California Berkeley and former lead of OpenAI's geopolitics team, told The Verge that "I think there are a lot of modifying words that are in the sentences that the [OpenAI] spokesperson gave." Shoker added that the vagueness of the language doesn't make it clear what exactly is prohibited here.
"The use of the word 'unconstrained,' the use of the word 'generalized,' 'open-ended' manner -- that's not a complete prohibition. That is language that's designed to allow optionality for the leadership … It allows leaders also not to lie to their employees in the event that the Pentagon does use the LLM in a legal manner without OpenAI leadership's knowledge." Based on what we've seen of OpenAI's existing contract and according to the Pentagon's current legal constraints, it could legally use OpenAI's technology to search foreign intelligence databases for information on Americans on a large scale.
So here we are. OpenAI has drawn its line in the sand, only to leave it blurry enough for a tank to roll through. The contract bans “bulk,” “open-ended,” and “generalized” data collection, but that’s not a prohibition; it’s a negotiation with the future.
As Shoker points out, those modifiers are escape hatches, not walls. The Pentagon can still legally run searches on foreign intelligence databases that sweep up Americans, at scale. The law hasn’t caught up.
Amodei knows it. Altman knows it. The fine print knows it.
And if the Pentagon’s policies shift tomorrow? OpenAI’s language is already designed to let that happen without anyone having to lie. What was sold as a red line looks more like a seam, one that can be stitched open whenever national security decides it needs a bigger net.
Related Reading
Browse all OpenAI news →
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic OpenAI researcher quits, citing distrust over ad‑driven engagement metrics
LLMs & Generative AI
Related Topic OpenAI launches GPT-Image 1.5 with precise editing for enterprise visuals
LLMs & Generative AI
From Archives Qualcomm's Elite chip targets AI wearables such as pendants, pins, and glasses
From Archives Lenovo unveils AI Workmate robot arm that scans, projects, and keeps company
From Archives OpenAI raises round larger than most tech firms, steps into Anthropic Pentagon void
Policy & Regulation
From Archives OpenAI CEO Sam Altman announces Pentagon deal with ambiguous safety principles
Common Questions Answered
What specific restrictions did OpenAI impose on U.S. government data processing?
OpenAI will no longer allow its AI models to ingest or process American data on a large-scale, unrestricted basis. The company specifically banned bulk, open-ended, or generalized data collection and analysis of U.S. citizens' information.
How does Dario Amodei view the current state of AI surveillance legislation?
Amodei has publicly warned that existing legislation has not yet caught up with AI's capacity for massive-scale surveillance and monitoring. He highlighted the significant gap between technological capabilities and current legal frameworks governing AI use.
What distinguishes OpenAI's approach from Anthropic's stance on government AI contracts?
While Anthropic was blacklisted for refusing to cross red lines around mass surveillance and autonomous lethal weapons, OpenAI chose to modify its contract to align with existing laws and Pentagon policies. OpenAI's approach appears to be more collaborative, seeking to work within current legal boundaries.
