Reddit Sues Perplexity, Others over Alleged Data Scraping (www.bloomberg.com)

Reddit has filed a lawsuit against Perplexity and other unnamed entities alleging large-scale scraping of its user-generated posts and comments to build and power AI services. The complaint claims the defendants bypassed Reddit’s controls and terms of service to copy public content en masse for training models and serving answers, harming Reddit’s business and users by misappropriating proprietary data and undermining platform controls. The suit centers on unauthorized collection, retention and commercial reuse of conversational data that Reddit says it did not license for AI training. The case is significant because it targets the core data pipelines feeding many large language models: publicly accessible social platforms with rich conversational datasets. A ruling for Reddit could force AI companies to shift from opportunistic scraping to negotiated licensing, clearer provenance tracking, and stricter compliance with platform rules and copyright law. Technically, it raises questions about dataset audits, attribution, opt-outs, and risk mitigation (e.g., differential privacy, filtering toxic content) when training models on scraped forum material. The outcome could reshape how models source high-value conversational data and accelerate industry moves toward licensed corpora or synthetic alternatives to avoid legal and ethical exposure.