Contract effective date: September 18, 2024. Runtime consent, request, security, audit, and deployment timestamps remain factual records.
Generative AI Content Safety and Moderation Policy
Scope
This policy applies to any feature that generates, transforms, summarizes, ranks or recommends text, image, audio, video, code, medical/health content, legal/financial content or user-generated content using AI.
Safety requirements
- Provide user reporting or flagging for problematic generated content when users can generate or publish content.
- Prevent generation or facilitation of CSAM, sexual exploitation, non-consensual intimate content, extremist content, hate, harassment, violence, self-harm encouragement, deceptive identity documents, malware, phishing, fraud, regulated goods instructions and unlawful activity.
- Label synthetic or AI-generated content where required by law, platform policy or risk assessment.
- Include human escalation for safety reports and high-impact output complaints.
- Maintain moderation logs with privacy safeguards and retention limits.
- Test prompts, jailbreaks and harmful content categories before release and after model changes.
- Do not represent AI outputs as professional advice unless licensed, validated and legally approved.
Mobile app requirements
The app must make report/flag controls visible in the AI output flow, not only in a web help center. Reports must feed safety evaluation and abuse-prevention improvements.