OpenAI · Usage Policies

Weapons and Critical Infrastructure Prohibition

High severity
Share 𝕏 Share in Share 🔒 PDF
Watch OpenAI Get alerts when this provision or policy changes.
Watch — $9.99/mo

Why it matters (compliance & risk perspective)

This provision applies an unusually high precautionary standard — prohibiting assistance even where the purpose is probably legitimate — which reflects the catastrophic and irreversible nature of mass-casualty weapon harm.

Consumer impact (what this means for users)

OpenAI's Usage Policy directly affects what you can ask ChatGPT or other OpenAI tools to do, and violations — even unintentional ones — can result in account suspension or termination without a guaranteed restoration process. The policy also makes developers and businesses that build on OpenAI's API responsible for ensuring their own users comply, meaning the apps you use powered by OpenAI may impose additional restrictions beyond OpenAI's own defaults. You can submit an appeal if OpenAI takes action against your account by visiting https://openai.com/transparency-and-content-moderation/#:~:text=determining%20enforcement%20actions.-,Appeals%20process.

How other platforms handle this

PayPal Medium

You may not use the PayPal service for activities that: violate any law, statute, ordinance, or regulation; relate to transactions involving (a) narcotics, steroids, certain controlled substances or other products that present a risk to consumer safety, (b) drug paraphernalia, (c) cigarettes, (d) it...

LinkedIn Medium

The Services are not for use by anyone under the age of 16. 'Minimum Age' means 16 years old. However, if law requires that you must be older in order for LinkedIn to lawfully provide the Services to you without parental consent (including using your personal data) then the Minimum Age is such older...

Spotify Medium

In any part of the Spotify Service, the Content that you access, including its selection and placement, may be influenced by commercial considerations, including Spotify's agreements with third parties. Some Content licensed by, provided to, created by, or otherwise made available by Spotify (e.g., ...

See all platforms with this clause type →

This clause could change without notice.

Get alerted when OpenAI updates this policy — with plain-language summaries and severity ratings.

Watch OpenAI Need compliance memos? Professional →
View original clause language
We don't want our models to be used to create content that could be used to cause real-world harm. This includes providing real uplift to people seeking to cause significant loss of life, e.g. those seeking to synthesize dangerous chemicals or bioweapons, even if it's probably for a legitimate reason (the risk simply being too high).

Applicable regulations

CFAA
United States Federal
DMCA
United States Federal
DSA
European Union

Provision details

Document information
Document
Usage Policies
Entity
OpenAI
Document last updated
March 5, 2026
Tracking information
First tracked
March 10, 2026
Last verified
April 27, 2026
Record ID
CA-P-002432
Document ID
CA-D-00005
Evidence Provenance
Source URL
Wayback Machine
SHA-256
d69a24617758e5b44e4be8eedeceb598a26dc4e280f2ab1469a45b64203e7403
Verified
✓ Snapshot stored   ✓ Change verified
How to Cite
ConductAtlas Policy Archive
Entity: OpenAI | Document: Usage Policies | Record: CA-P-002432
Captured: 2026-03-10 03:28:59 UTC | SHA-256: d69a24617758e5b4…
URL: https://conductatlas.com/platform/openai/usage-policies/weapons-and-critical-infrastructure-prohibition/
Accessed: May 4, 2026
Classification
Severity
High
Categories

Other risks in this policy

Don't miss changes to this clause.

OpenAI has updated this policy before. Get alerted on the next change.

Watch OpenAI

Frequently Asked Questions

What does OpenAI's Weapons and Critical Infrastructure Prohibition clause do?

This provision applies an unusually high precautionary standard — prohibiting assistance even where the purpose is probably legitimate — which reflects the catastrophic and irreversible nature of mass-casualty weapon harm.

Is ConductAtlas affiliated with OpenAI?

No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by OpenAI.