📰 Key Takeaways

Meta recently updated its commercial terms to add a clause stating the company can “use any content on a business account’s domain, web pages, or URLs to provide, develop, improve, research, train, and maintain” the integrity of Meta’s products and services. Legal experts interviewed by Nieman Lab called the clause “unusual,” saying it effectively authorizes Meta to scrape the website content of any publisher with a Meta business account and use it to train the company’s AI models. Notably, Meta already won a high-profile copyright infringement lawsuit last year, though other lawsuits against AI companies (including Meta) are still ongoing.

However, a Meta spokesperson insisted that the updated terms didn’t reflect any change in how the company scrapes web content. The Verge then reported that the controversial clause had been removed. In a statement, Meta said: “We heard the feedback and want to clarify: the consolidated scraping language we added only applies to our ads and business products, and we’ve clarified that in the terms themselves. This doesn’t change how Meta uses web crawlers.”

Overall, this episode reflects the ongoing legal and public pressure AI companies face over content licensing and training data sourcing — even small wording changes in terms of service can be read as expanding data-collection rights, sparking controversy and public backlash.

💬 JudyAI Lab’s Take

We think this quick update-then-walkback of Meta’s commercial terms makes a neat case study of the AI industry’s ongoing data-licensing disputes.

The incident itself is simple: Meta added a line to its commercial terms saying it “can use any content on a business account’s domain or web pages to train Meta products,” and legal experts argued this effectively authorized scraping the websites of all Meta business account customers. Once the controversy broke, the clause was quickly pulled, and Meta clarified that the consolidated scraping language only applies to ads and business products — and doesn’t change its existing crawling practices.

For AI builders, this highlights a persistent tension: the boundaries of training data acquisition are often buried in the fine print of terms of service, not announced in public policy. When a product and AI training are bundled into the same terms, users have a hard time telling how broad the scope of their consent actually is. This episode also shows that public and legal scrutiny can genuinely get terms corrected in short order.

Next time you’re about to sign off on any platform’s commercial terms, it’s worth spending a few minutes reading through the data-use section to make sure the scope of your consent isn’t broader than you expected.

📅 Source Info

🔗 Further Reading