What OpenAI publishes about its security, privacy and use of your data, with a link to the page each fact came from. Checked November 22, 2026.
"Not found" means our reader didn't find it on the public pages it could read; the vendor may still publish it.
Statements differ between pages
OpenAI's own words (openai.com/security-and-privacy):
You decide whether your data is used for training and model improvement.
Vendors often have different rules for consumer or free plans and for business or enterprise plans. Check the terms for the plan you would use.
If you do not want us to use your Content to train our models, you can opt out by following the instructions in this article.
openai.com/policies/terms-of-use
Read our instructions(opens in a new window) on how you can opt out of our use of your Content to train our models.
openai.com/policies/privacy-policy
These include the following controls: You can easily choose whether your Content can be used to improve and train our models(opens in a new window).
openai.com/policies/privacy-policy
Different pages say different things. This is often a split between business plans and free or consumer plans. Read each quote for the plan you would use.
| Certification or report | Mentioned on |
|---|---|
| SOC 2 Type II | openai.com/security-and-privacy |
| ISO/IEC 27001 | openai.com/security-and-privacy |
| ISO/IEC 27701 | openai.com/security-and-privacy |
| ISO/IEC 42001 | openai.com/security-and-privacy |
| HIPAA | openai.com/security-and-privacy |
| PCI DSS | openai.com/security-and-privacy |
| CSA STAR | openai.com/security-and-privacy |
Privacy laws and frameworks mentioned: GDPR, CCPA/CPRA. These are laws the vendor says it follows, not certifications.
A certification on a website is a claim. For anything sensitive, ask OpenAI for the SOC 2 report or ISO certificate.
Our reader didn't find a subprocessor list on OpenAI's public pages. It may be published somewhere we couldn't reach, such as a trust portal that needs JavaScript or a login. Ask OpenAI for it if you need it.
| SSO / SAML | Not found |
| SCIM provisioning | Not found |
| Multi-factor authentication | Not found |
| Encryption at rest | Mentioned |
| Encryption in transit | Mentioned |
| Penetration testing | Not found |
| Bug bounty / disclosure program | Mentioned |
| Audit logs | Mentioned |
| Data residency | Mentioned |
| security.txt contact | https://bugcrowd.com/openai |
| Trust / security center | openai.com/security-and-privacy |
| Privacy policy | openai.com/policies/privacy-policy |
| Subprocessor list | Not found |
| Data processing agreement (DPA) | Not found |
| AI policy | Not found |
| Terms | openai.com/policies/terms-of-use |
| Status page | status.openai.com |
| Check | Points | Result |
|---|---|---|
| Security or trust page | 15 | Yes |
| Certifications listed | 20 | Yes |
| Privacy policy | 10 | Yes |
| Subprocessor list | 15 | No |
| DPA available | 10 | Yes |
| AI training policy stated | 15 | Yes |
| security.txt contact | 5 | Yes |
| Status page | 10 | Yes |
The score measures how much of a standard security review our reader could answer from OpenAI's public pages. It is not a rating of how secure OpenAI is, and a "No" can mean the document exists where our reader couldn't see it.
Get the same profile for any vendor in seconds, or let your AI agent do it with Stormap's free MCP server.
Anthropic · Cohere · Copy.ai · Descript · ElevenLabs · Glean · Harvey · Hugging Face · Jasper · Make · Mistral AI · Pinecone
Source: OpenAI's own public website (4 pages read on November 22, 2026). Stormap is not affiliated with OpenAI. Facts are what the vendor publishes about itself; certifications are claims, not verified audit reports. Something wrong or out of date? Pages are re-checked every two weeks.