Meta's AI can still scrape news sites
What's the story
Meta's revised business terms for its business accounts have come under fire from legal experts. The changes, which go into effect on October 30, were initially thought to give the company a free hand in using content from news sites for AI training. However, despite the removal of explicit language around "train" its products with scraped content, concerns remain about their potential implications.
Expert opinions
Legal experts weigh in on potential implications
Legal experts, including James Grimmelmann from Cornell Law School, have raised concerns over the new terms.
He said, "This does nothing to restrict the 'associated with your products or services' [language] that would let Meta scrape entire sites."
The initial version of these terms gave Meta permission to use any content from websites linked with business accounts for developing and training its products.
Scraping concerns
'Develop' language raises further concerns
Despite the backlash from the publishing industry, Meta's revised terms still permit the use of scraped content for developing any "business product."
Matt Topic, a media and intellectual property attorney at Loevy & Loevy, told Nieman Lab that he thinks "I think 'develop' is broad enough to include 'train.' "
This further fuels concerns over Meta's intentions with its AI models and their potential impact on news sites.
Company response
Meta insists update not about scraping news sites
Meta has defended its updated terms, insisting they are not meant to scrape news content for free.
In a statement to The Verge, Meta spokesperson Daniel Roberts said "content owners can use robots.txt to signal their preferences."
He clarified that the update "is not about scraping news sites," and these changes don't give them new rights that would override those preferences.