AI Scraping Concerns
Meta has modified its business terms to remove phrases implying it could freely train artificial intelligence models on publisher websites, though legal professionals argue the adjustments fall short of resolving fundamental concerns.
Meta’s AI could still scrape news sites, legal experts say
The updated terms may still provide justification for utilizing scraped web data in artificial intelligence development.
Terms and Block Instructions
Initially, the updated business account terms appeared to grant the company authority to use website content for research and model training. Legal analysts previously noted this could potentially permit the company to argue that commercial agreements override standard crawling restrictions, such as robots.txt directives.
Revised Language and Loopholes
Following industry criticism, the explicit reference to training products was removed. However, specialists suggest terms allowing content use for developing business products remain broad enough to encompass model training activities.
Company Position
The company has stated that the revisions are not intended to facilitate uncompensated news scraping, reiterating that site operators retain the ability to manage crawling preferences through standard protocol files.
