OpenAI Usage Policies 2025-10-29

Citation

OpenAI. “Usage policies.” Policy web page, last changelog entry 2025-10-29. https://openai.com/policies/usage-policies/. Read 2026-10-15.

Bottom line

• OpenAI has one set of usage policies for all of its products. The policies have four groups: “Protect people”, “Respect privacy”, “Keep minors safe” and “Empower people”.
• The policies prohibit high-stakes decisions in 13 sensitive areas “without human review”.
• OpenAI requires its “review and approval” for national security or intelligence use. This is an approval-based tier.

Background

This page is one row of that table.

The Oct. 29, 2025 update made “a universal set of policies across OpenAI products and services.” (Changelog) The page that the web reader returned shows no separate effective date.

Key claims

Introduction:

• “We hold people accountable for inappropriate use of our services, and breaking or circumventing our rules and safeguards may mean you lose access to our systems or experience other penalties.”
• “We monitor and enforce policies with privacy safeguards in place and clear review processes.”
• “You can appeal if you think we have made a mistake enforcing policy, and we will work to make things right.”

“Protect people”. “You cannot use our services for”:

• “weapons development, procurement, or use, including conventional weapons or CBRNE”
• “destruction, compromise, or breach of another’s system or property, including malicious or abusive cyber activity”
• “threats, intimidation, harassment, or defamation”
• “terrorism or violence, including hate-based violence”
• “provision of tailored advice that requires a license, such as legal or medical advice, without appropriate involvement by a licensed professional”
• “unsolicited safety testing”
• “circumventing our safeguards”
• “national security or intelligence purposes without our review and approval”

“Respect privacy”:

• “we don’t allow attempts to compromise the privacy of others, including to aggregate, monitor, profile, or distribute individuals’ private or sensitive information without their authorization.”
• “facial recognition databases without data subject consent”
• “real-time remote biometric identification in public spaces”
• “evaluation or classification of individuals based on their social behavior, personal traits, or biometric data (including social scoring, profiling, or inferring sensitive attributes)”
• “assessment or prediction of the risk of an individual committing a criminal offense based solely on their personal traits or on profiling”

“Keep minors safe”:

• “child sexual abuse material (CSAM), whether or not any portion is AI generated”
• “grooming of minors”
• “We report apparent child sexual abuse material and child endangerment to the National Center for Missing and Exploited Children.”

“Empower people”:

• “deceit, fraud, scams, spam, or impersonation”
• “political campaigning, lobbying, foreign or domestic election interference, or demobilization activities”
• “automation of high-stakes decisions in sensitive areas without human review”. The policy lists 13 areas:
  • critical infrastructure, education, housing, employment
  • financial activities and credit, insurance, legal, medical
  • essential government services, product safety components, national security, migration, law enforcement

Evidence and method

This is a policy document, not research. It states rules. It gives no data on how often OpenAI enforces them.

Limits

• openai.com refused the download (HTTP 403). The copy comes from a web reader, so the live page is the authority. The copy can differ from the live page in layout.
• The copy has no “effective” date line. The date in the title is the latest changelog entry.
• The policies list categories only. They do not define terms such as “high-stakes decisions” or “human review”.
• The page does not name a research exception. The policy prohibits “unsolicited safety testing”. Thus solicited testing can be possible, but the page does not say how to ask for it.
• The policies do not name critical infrastructure as a prohibited use. Critical infrastructure appears only in the list of areas that need human review.