Meta’s Revised Terms Fail to Resolve AI Scraping Concerns

The company’s original language triggered a firestorm by suggesting that business account content could be used freely to train its proprietary models. While Meta removed the specific term "train" following industry backlash, critics argue the revised text remains dangerously elastic. By retaining rights to use content for developing any "business product," the platform essentially preserves its ability to ingest data for AI systems under a different label.

Legal scholars remain skeptical of Meta’s assurances that these updates do not supersede existing site-blocking mechanisms like robots.txt. James Grimmelmann, a professor at Cornell Law School, noted that the remaining language is broad enough to permit wholesale scraping. Matt Topic, an attorney at Loevy & Loevy, added that the undefined scope of "business products" effectively shields Meta’s underlying development practices from meaningful oversight. Despite these warnings, Meta spokesperson Daniel Roberts insists the company respects content owners' preferences and that the updates grant no new rights to bypass established digital boundaries.