AI Video Compliance and Content Safety Review System

You are a content compliance engineer. Build an automated video compliance review system covering regulatory, platform, brand, and legal requirements before release. Inputs: - Content library: [INSERT - UGC platform, advertisements, films, corporate video] - Ruleset: [INSERT - regional regulations, platform policies, brand safety, age rating] - Languages and markets: [INSERT] - Review SLA and escalation path: [INSERT] - Human review team capacity: [INSERT] Deliver in modules: 1. Objective: detect violations in visuals, speech, on-screen text, and music with auditable decisions. 2. Process: multimodal scan (violence/NSFW/self-harm/hate speech, weapons, drugs), ASR/OCR for banned claims and missing disclaimers, logo/music rights and political-sensitivity checks -> risk scoring -> human review queue -> policy decision logging. 3. Toolchain: [INSERT - moderation models, ASR/OCR, rules engine, case-management platform], hash/blacklist matching. 4. QA: precision/recall against a labeled set, false-appeal rate, red-team tests, drift monitoring, and policy versioning. 5. Delivery: decision reports with timestamps and evidence frames, reviewer console, audit trail, regulator-ready exports. Output: rule taxonomy, risk matrix, model/vendor selection, and market-by-market coverage table. ================