Meta has quietly removed language from its policies that raised concerns about the company's ability to scrape news websites and use their content to train artificial intelligence models. The change comes amid ongoing tensions between tech giants and news publishers over AI training data.
The removed language had suggested Meta could collect content from news sites without explicit permission, sparking backlash from media organizations. Publishers argued that their journalistic content shouldn't be used to train AI systems without compensation or consent.
The policy revision signals a shift in Meta's approach to content acquisition for AI development. While the company continues to build large language models and generative AI features, the removal suggests recognition that scraping news content without clear authorization creates legal and reputational risks.
Industry observers note this move may pressure other tech companies to clarify their own data collection practices. News organizations have increasingly demanded explicit licensing agreements rather than relying on vague terms of service that could be interpreted to allow widespread scraping.
The change doesn't necessarily mean Meta has stopped using news content for AI training, but it does indicate the company is reconsidering how it acquires training data in an environment of growing scrutiny over AI ethics and copyright.
