Key points
- Enterprise customers will bear legal liability for mistakes made by Meta’s AI models under terms effective October 30.
- Meta grants itself broad rights to use automated tools to access and analyze content associated with customer accounts.
- The company also holds customers accountable for actions taken by third-party partners or contractors using their services.
Meta is shifting legal responsibility for artificial intelligence errors from itself to enterprise customers in updated commercial terms scheduled to take effect on October 30. These changes also expand the company’s authority to scrape web content linked to business accounts and hold organizations accountable for the actions of their contractors.
How it unfolded
- On Monday, Meta published new commercial terms on its website that shift AI liability to enterprise users and expand data collection rights.
- Consultants identified web scraping, data ownership, and liability transfer as the three primary concerns regarding the proposed language changes.
- Late Monday, Meta public affairs manager Emil Vazques emailed a statement suggesting the company was rethinking its web scraping position.
- Vazques clarified that crawling terms apply only to ads and business products, and that robots.txt preferences remain respected, though the document had not yet been updated as of Monday evening.
Who is affected
Enterprise customers using Meta’s business products will face new legal obligations under the October 30 terms. These organizations will be held liable for any mistakes generated by Meta’s AI models. The language also extends this responsibility to cover actions performed by partners or contractors acting on behalf of the customer. This creates a broader scope of accountability for IT executives managing these commercial relationships.
What happens next
The proposed terms are set to become effective on October 30. As of Monday evening, the published document had not yet been altered to reflect Meta’s clarification regarding web scraping, though the company indicated a change was coming soon. The updated language aims to clarify that crawling is limited to advertising and business product maintenance, not general news site scraping, according to Emil Vazques.
The background in plain language
Liability shifts determine who bears the legal and financial cost when technology fails or causes harm. In this context, transferring liability means the customer accepts risk for AI outputs they did not directly control. Web scraping involves automated tools collecting data from websites, often bypassing standard user interactions. Robots.txt is a standard file that website owners use to signal which areas of their site should not be accessed by automated crawlers. Understanding these concepts helps teams evaluate the operational and legal risks of new vendor terms.
What to do and how to stay safe: Meta
- Review the updated commercial terms carefully to understand the specific scope of liability for AI-generated content and partner activities.
- Verify whether your organization’s web properties are associated with Meta business accounts that might be subject to the new crawling permissions.
- Consult with legal counsel to assess the impact of accepting responsibility for third-party contractor actions under these new agreements.
- Monitor Meta’s official channels for the updated document text that reflects the clarified stance on robots.txt and web scraping limitations.
Step-by-step guide: Injection attacks examples: how prompt leaks break AI systems
General security guidance from the Circuit Observer newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
When do Meta's new commercial terms take effect?
The new commercial terms are scheduled to kick in on October 30.
Does Meta still respect robots.txt files?
According to Meta public affairs manager Emil Vazques, content owners can still use robots.txt to signal their preferences, and the updates do not supersede those signals.
Who is liable for AI mistakes under the new terms?
The new language shifts the liability for mistakes made by Meta’s AI models to its enterprise customers.
