Authorized use
The Terms of Service restricts illegal activity, infringement, deceptive synthetic media, fraudulent content, unauthorized access, system disruption, and reverse engineering. It also requires users not to submit data they do not have rights to process.
Decision use
A detection score or label is not described as proof. The Terms of Service calls the output probabilistic and identifies human review as necessary for consequential use where law or internal policy requires it. Customers remain responsible for defining the review, escalation, and decision process around an output.
Evidence boundary
Limitations
- This page is not a separate AI policy and does not create permissions, warranties, or model-performance commitments.
Source record