Meta has quietly removed a controversial provision from its commercial terms of service that legal experts warned could have allowed the tech giant to scrape the websites of publishers and businesses to train its artificial intelligence models. The swift reversal follows widespread concern within the media and legal communities over broad phrasing that appeared to grant the company sweeping rights to utilize external web content for product development and AI research.
The controversy began when Meta updated its commercial terms of use, introducing language stipulating that the company could use any content originating from a business account’s domains, web pages, or URLs for a multitude of purposes. Specifically, the updated terms stated that the data could be leveraged to provide, develop, improve, research, train, and maintain the integrity of Meta products and services.
According to legal experts interviewed by Nieman Lab, this development looked like a remarkable provision that could give Meta explicit permission to scrape the website of any publisher maintaining a Meta business account. The fear was that routine commercial interactions or business page setups would inadvertently sign away publishers’ digital rights, allowing the tech conglomerate to feed proprietary news articles, investigative reporting, and copyrighted journalistic works directly into its proprietary AI training pipelines.
The anxiety surrounding Meta’s updated terms arrived against a backdrop of escalating legal friction between media publishers and artificial intelligence developers. The intersection of copyright law and generative artificial intelligence has become one of the most fiercely contested battlegrounds in the technology sector. While Meta successfully defeated a high-profile lawsuit last year when a federal judge sided with the company in a legal challenge over training AI models on copyrighted books, the broader legal war is far from over. Numerous other lawsuits filed by authors, artists, and news publishers against various AI companies, including Meta, continue to wind their way through the court system, with plaintiffs arguing that unauthorized scraping and ingestion of protected works constitute widespread copyright infringement.
Despite the alarm raised by legal analysts, a Meta spokesperson initially insisted that the updated terms did not reflect a change in the way Meta crawls the web or handles external publisher content. The company maintained that the language was merely a consolidation of existing policies rather than a newly minted grab for publisher data. However, as scrutiny mounted and industry watchdogs highlighted the potential implications of the policy, the situation evolved rapidly.
Reporting from The Verge confirmed that Meta has officially removed the contested language from its terms of service, eliminating the specific phrasing that had stoked fears of aggressive data harvesting from publisher domains.
In a subsequent statement addressing the removal, the company sought to quell industry anxieties and clarify the intended scope of the policy. "We’ve heard you and want to clarify: the crawling terms we consolidated only applied to our ads and business products, and we’ve clarified that in the terms themselves," the company said in its statement. "It does not change how Meta uses web crawlers."
The rapid inclusion and subsequent deletion of the provision underscore the intense sensitivity surrounding AI training data, intellectual property rights, and the relationship between major technology platforms and news publishers. As legal battles continue to define the boundaries of what tech companies can and cannot do with publicly accessible web content, publishers remain hyper-vigilant regarding any shifts in platform terms that could impact their digital archives and journalistic output.
