Meta Revises Its Terms, Lawyers Remain Concerned
Meta has softened the wording of its commercial terms by removing the explicit mention of 'training' its products with publisher content. The company claims it does not seek to override robots.txt, but legal experts believe the revised text could still allow for scraping and use for AI.
Meta cites robots.txt and denies any new rights
Meta asserts that its new terms are not intended to extract news content for free. Its spokesperson, Daniel Roberts, states that content owners can express their preferences through the robots.txt file. He clarifies that the update is not related to scraping news sites and does not grant them new rights that would override these preferences. Historically, robots.txt serves to inform scrapers, including search engines, about what they are allowed to crawl.
Legal experts point out flaws in the revised version
Several legal specialists believe that the changes do not alleviate their concerns. James Grimmelmann argues that the retention of the phrase "associated with your products or services" would still allow Meta the possibility to scrape entire sites. Attorney Matt Topic notes that the revised version would still permit the use of extracted content to develop any "commercial product," a term he sees as undefined and which he believes would encompass AI products. He also considers that the verb "develop" is broad enough to include the idea of "training." Other legal experts had already indicated that these terms could be invoked to override blocking directives like robots.txt.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
What the terms say and the effective date
Meta has modified its commercial terms by removing the explicit reference to "training" its products with collected content, following a strong reaction from the publishing industry. A previous version stated that the company could use any content from sites associated with professional accounts to develop, research, and train its products. These formulations had been interpreted as granting permission to train AI on news sites. The effective date for these terms is set for October 30.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.