Filter streaming catalogs to remove content by accused/convicted actors

Many streaming customers—from parents and survivors to ethically-motivated viewers—have no reliable way to avoid titles because they feature accused or convicted actors, and platforms face recurring PR and churn risks when those titles surface; this is a clear pain point against a large addressable market of roughly 1.10 billion paid streaming subscribers and an estimated $52.7B in annual paid streaming spend (1.10B x $3.99/mo), with market and revenue potential scores of 90/100 and 88/100 respectively. You could build a multi-tier product: a catalog-indexing engine that combines credits metadata, AI-based face recognition and entity-resolution to tag and score titles, consumer-facing filters (browser/smart-TV extension and mobile app) and a B2B API for platforms and aggregators. Start with a metadata-first approach augmented by AI verification and human review for edge cases, and commercialize via platform licensing and a premium consumer tier; the initial technical work is to index English-language catalogs (a feasible MVP of ~50k titles) and validate accuracy. This timing is favorable because streaming personalization and ethical-consumption trends are accelerating and advances in AI vision/entity-resolution make automated identification practical, but the project has real execution risks: compute and indexing costs, name-disambiguation, false positives, and significant legal/defamation and privacy considerations when labeling people as “accused” versus “convicted.” To stand out you must prioritize legal defensibility and high precision—secure access to credits/metadata to reduce reliance on face models, provide transparent controls and appeals, and pursue B2B partnerships first to capture higher ARPU while proving efficacy in a limited pilot; if those risks are managed, the idea is commercially promising but requires cautious, evidence-driven execution.