Content ModerationInternal OnlyVerified
Reviewed and published by trentmaziarz, March 24, 2026. Discovered and drafted by our automated research pipeline.
OpenAI uses internal and external AI safety evaluations — including red-teaming — to test its models for dangerous capabilities before and after deployment.
Details
OpenAI's Preparedness Framework requires safety evaluations of all frontier models before deployment. The company runs an external Red Teaming Network of independent experts who probe models for risks including CBRN (chemical, biological, radiological, nuclear) threats, cybersecurity, and persuasion. OpenAI and Anthropic conducted a joint safety evaluation in 2024. Evaluations cover models including GPT-5, o-series, Sora, and Operator.
Products affected
All frontier modelsGPT-4GPT-4oGPT-4.5GPT-5o1o3SoraOperator
Sources & Evidence
News Article
Company Disclosure
Cite this record
Trace Foundation. (2026). OpenAI: OpenAI uses internal and external AI safety evaluations — including red-teaming — to test its models for dangerous capabilities before and after deployment (data as of 2026-07-22) [Data set record]. AI Trace. https://www.aitrace.org/r/practice/0f9cfd4a-ed3c-4074-88b3-ad202638b3b8. Accessed October 14, 2026.
- Stable link
- https://www.aitrace.org/r/practice/0f9cfd4a-ed3c-4074-88b3-ad202638b3b8
- Data as of
- July 22, 2026
- Last verified
- April 10, 2026
Other practices by OpenAI
Creative GenOpenAI offered Sora, an AI video generation tool that produced short video clips from text prompts or uploaded images, to ChatGPT Plus and Pro subscribers starting December 2024; the product was discontinued in a phased shutdown, with the app and web closing on April 26, 2026 and the API scheduled to close on September 24, 2026. OpenAI announced the discontinuation on March 24, 2026, with reporting pointing to high compute costs and a strategic shift toward coding and enterprise products. A three-year licensing deal signed in December 2025 with The Walt Disney Company, under which Disney had committed a $1 billion equity investment in OpenAI, lapsed when the product shut down before the partnership terms were executed.OtherOpenAI deployed ChatGPT for US government and defense use through its OpenAI for Government platform, including a $200 million Pentagon contract and deployment on the DoD's GenAI.mil platform for approximately 3 million personnel.RecommendationOpenAI deployed contextual ads in ChatGPT starting February 9, 2026, showing sponsored content to free-tier users based on the topics of their current conversation.
Have evidence about OpenAI's AI practices? Submit a report.
Report a Sighting →