Meta’s Revised Terms Still Allow Broad Scraping Despite Removing AI Training Language

Key points

• Meta removed the word 'train' from its updated terms of service on October 17, 2026, following criticism that the original draft allowed scraping content to train AI models.
• The revised terms limit scraped content usage to 'advertising and business products' rather than all Meta products, but legal experts note 'business products' is undefined and could encompass AI tools.
• The new terms allow Meta to use automated software to scrape websites associated with business accounts, potentially overriding robots.txt files, which are not legally enforceable.
• The changes take effect on October 30, 2026, and there is no option for business accounts to opt out; the only alternative is to cease using Meta's platforms.
• Meta claims the update merely clarifies that crawling applies to ads and business products, but critics argue the language remains too broad and ignores standard web scraping preferences.

Background

This development follows earlier reporting in October 2026 that Meta’s commercial terms would grant broad scraping rights to business websites, sparking concerns among news publishers and legal experts. Previous coverage highlighted that the initial terms explicitly mentioned training AI models, leading to immediate backlash from media trade associations and regulators.

How outlets are covering it

Nieman Lab and legal experts like Matt Topic and James Grimmelmann argue that the revised terms remain problematic because 'business products' is undefined and could include AI training. The Register’s Eric Goldman emphasizes that Meta’s history of privacy issues makes good-faith interpretation unlikely, noting that robots.txt compliance is not legally required. In contrast, Meta’s spokesperson claims the changes only clarify existing policies and that publishers can still use robots.txt to block crawling, though this is not explicitly stated in the new terms. The News/Media Alliance calls the terms 'outrageous and probably illegal,' highlighting the lack of opt-out options.

Why it matters

The outcome affects how businesses, especially news publishers, can protect their content from being scraped by Meta for AI training or other commercial purposes. If the terms are enforced as written, companies may lose control over their digital assets without legal recourse, setting a precedent for platform power over content creators.

What to watch

The revised terms take effect on October 30, 2026. Regulators in the EU and UK may intervene, as the European Publishers Council and UK News Media Association have called for scrutiny. Businesses may need to review their Meta account usage or seek legal advice to protect their content.