Allowed uses
- Processing your own authorized AI-generated image for editing, interoperability, research, or testing.
- Removing ordinary metadata when you have permission to alter the file.
- Using provider checkers for content review, labeling, fact-checking, or trust and safety.
Prohibited uses
- Removing visible stock-photo, photographer, publisher, rights-holder, or ownership watermarks.
- Misrepresenting AI-generated or modified content as human-created where that is deceptive or disclosure is required.
- Fraud, impersonation, harassment, election manipulation, evidence falsification, or evasion of platform safety enforcement.
- Automating repeated detector checks to reverse-engineer or tune watermark evasion.
- Bypassing quotas, attacking infrastructure, or accessing another user’s files or tokens.
Enforcement
Jobs may be blocked, canceled, or deleted when use appears inconsistent with this policy or provider requirements.