Media & Culture

Meta's AI Terms Still Worry Publishers Over Your Website Content

Meta's AI Terms Still Worry Publishers Over Your Website Content

⚑Your online content might be feeding Meta's AI without your explicit permission.

Deep Dive

Meta recently adjusted its new terms for business accounts, which originally seemed to give them broad permission to train AI (AI learning from data) using content from associated websites. After strong reactions from publishers, Meta removed the explicit word "train." However, legal experts remain concerned, arguing that the updated terms are still broad enough for Meta to "develop" AI products using content "scraped" (automatically copied) from your site, potentially without your full consent. This means if you run a business with an online presence or publish content, your work could be contributing to Meta's AI models, raising questions about content ownership and fair use.

The core of the concern lies in Meta's "develop" clause for business products. Attorneys like Matt Topic argue that "develop" is broad enough to include training AI, especially when "business products" likely refer to their AI tools. This could mean that even if you've set up a "robots.txt" file (a simple instruction telling automated tools what parts of your website they can access or copy), Meta might argue their business terms take precedence. For content creators and news organizations, this is a significant issue. It could mean Meta profits from your hard work without direct compensation or clear permission, undermining traditional models for online publishing.

Meta insists these changes aren't about scraping news sites and that "content owners can use robots.txt to signal their preferences," adding that the updates don't grant them new rights to override these preferences. Yet, experts like James Grimmelmann from Cornell Law argue the language around "associated with your products or services" is still too vague and could allow wide-ranging scraping. This ongoing ambiguity means that while Meta *says* one thing, their legal terms might enable another, leaving publishers and businesses in a gray area regarding their intellectual property and how their content fuels the next generation of AI tools.

Key Points
  • Meta changed its AI terms for businesses, but experts still worry about content scraping (automated copying of website data).
  • Legal experts, including James Grimmelmann from Cornell Law School, say the updated terms remain too broad.
  • This could mean your website content helps train Meta's AI products, potentially affecting your control and business model.

Why It Matters

Your online content or business website could be used to build Meta's AI, impacting ownership and fair compensation.

πŸ“¬ Get the top 10 AI stories daily