Meta Revises Commercial Terms Following Scraping and AI Training Concerns

Meta has updated its commercial terms of use to remove a provision that raised concerns about web scraping and artificial intelligence training. The company clarified that the language in question only applied to its advertisements and business products, according to a statement. This change addresses feedback regarding the scope of content usage permissions granted to the social media giant.

Root cause

The initial update to the commercial terms included language declaring that Meta could use content from business account domains, web pages, or URLs. The text stated this content could be used to provide, develop, improve, research, train, and maintain the integrity of Meta products and services. Legal experts interviewed by Nieman Lab described this as a remarkable provision. They argued it could effectively give Meta permission to scrape the website of any publisher holding a Meta business account. This raised questions about whether such content could be used to train the company’s AI models without additional consent.

Attack path

While no technical attack occurred, the legal and operational risk involved the automatic ingestion of web content. The provision suggested that by maintaining a business account, publishers implicitly allowed Meta to crawl their sites. This mechanism could have been interpreted as a broad licence to harvest public-facing content for machine learning purposes. The concern was not about hacking or intrusion, but about the contractual basis for data collection and subsequent use in AI development, potentially impacting copyright considerations.

Affected versions

The provision affected the latest version of Meta’s commercial terms of use at the time of the update. It applied specifically to entities holding Meta business accounts. The company has since removed the controversial language from the terms. The update does not reflect a change in how Meta crawls the web, according to the company. The change is limited to the written terms governing business relationships rather than technical infrastructure.

Mitigation

• Review current commercial terms of service for any new data usage clauses.
• Verify whether business account agreements include broad content licensing permissions.
• Consult legal counsel regarding implications for copyright and intellectual property rights.
• Monitor vendor announcements for clarifications on web-crawling and AI training practices.

Why it matters

Teams usually rely on clear contractual boundaries to understand data usage rights. When vendors update terms, it is essential to distinguish between technical changes and legal permissions. In this case, the concern was about the scope of content licensing for AI training. Organisations should review how their content is defined in service agreements. This helps ensure that automated data collection aligns with corporate policy. Understanding the difference between operational crawling and contractual permission is key to managing legal risk in digital partnerships.

What to do and how to stay safe: Meta

• Audit existing business agreements with major tech platforms for broad content usage clauses.
• Ensure that internal policies align with the latest terms of service regarding data rights.
• Monitor for further clarifications from vendors on how user content is utilised in AI development.
• Engage legal teams to assess potential copyright implications of default business account terms.

Step-by-step guide: Stop Sensitive Data Leaks in AI Chatbots: Priority Controls

General security guidance from the Vector Bulletin newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Did Meta change how it crawls the web?

No, a Meta spokesperson insisted that the updated terms did not reflect a change in the way the company crawls the web.

What did the removed provision allow?

It declared Meta could use content from business account domains to develop, train, and maintain its products and services.

Why was the provision removed?

Meta removed the language after hearing concerns, clarifying that the terms only applied to ads and business products.