Meta’s AI Could Still Scrape News Sites Despite New Terms, Legal Experts Warn
Meta’s updated business terms may still provide legal cover for training its AI models on scraped news content, according to legal experts. The revised language, they argue, leaves room for the company to justify using third-party sites for AI training—raising fresh concerns for publishers already wary of unauthorized data harvesting.
The Loophole in Meta’s New Terms
Meta recently introduced changes to its business terms that ostensibly clarify how user data and third-party content are handled. However, legal analysts say the updated language is ambiguous enough that Meta could still claim implicit permission to scrape news sites for AI training. The terms do not explicitly prohibit automated data collection, nor do they require explicit consent from publishers, leaving a gap that could be exploited.
“The terms are written in a way that gives Meta considerable flexibility,” said one intellectual property lawyer familiar with the matter. “They don’t outright authorize scraping, but they don’t rule it out either. In a legal dispute, Meta could point to the absence of a clear prohibition as justification.”
Publishers’ Concerns Mount
News organizations have long objected to AI companies using their content without licensing agreements or compensation. While some major publishers have struck deals with AI developers, many smaller outlets lack the leverage to negotiate. Meta’s latest terms, critics say, exacerbate the power imbalance by making it harder for publishers to assert control over their content.
“This is a pattern we’ve seen across the industry,” said a media rights advocate. “AI companies draft terms that are vague enough to allow scraping while maintaining plausible deniability. Publishers are left to play catch-up in court.”
Regulatory Landscape in 2026
As of 2026, the regulatory environment for AI training data remains patchy. The European Union’s AI Act has begun to enforce transparency requirements, but enforcement across borders is inconsistent. In the United States, federal legislation on AI and copyright has stalled, leaving individual states and courts to interpret existing laws. This fragmented landscape makes it difficult for publishers to mount a unified legal challenge.
Meanwhile, Meta has continued to expand its AI capabilities, recently unveiling new models that require vast amounts of data. The company has not publicly commented on the specific allegations, but has previously stated that it respects intellectual property rights and complies with applicable laws.
What’s Next?
Legal experts suggest that publishers may need to rely on contractual terms, technological measures like paywalls and bot-blocking, or collective bargaining to protect their content. Some are calling for clearer legislation that explicitly addresses AI scraping.
“Without a definitive legal ruling or new statute, the ambiguity will persist,” said another legal analyst. “Meta and other AI firms will continue to operate in the gray area until forced to change.”
For now, the debate underscores the growing tension between AI innovation and content ownership—a conflict likely to intensify as AI models become more data-hungry.
via The Verge AI