Meta to scrape business customer websites for AI model training
Meta plans to update commercial terms to allow automated scraping of business-linked domains for AI training, drawing sharp criticism from media groups.
Key points
- Meta will authorize automated software to collect content from domains associated with its business customers.
- The new terms allow the use of scraped data to train AI models and improve Meta products.
- Industry leaders, including the News/Media Alliance, have condemned the move as harmful to publishers.
Meta intends to introduce new commercial terms at the end of the month that would authorize the company to scrape content from websites associated with its business customers. According to a report by The Register, this change allows Meta to use automated software to collect, analyze, and retrieve data from these domains for the purpose of training artificial intelligence models or improving its existing products.
The updated language explicitly grants Meta the right to create test credentials and use web scrapers, bots, and crawlers to access any domains, web pages, or URLs linked to a business customer’s services. Meta states it may use any content retrieved from these sources to research, develop, and maintain the integrity of its products, including AI systems.
Scope of the Policy Change
Meta defines business customers broadly for this purpose, including any entity using ads, selling products, developing apps, or managing a Page or Group for business purposes. This definition covers users of widely adopted services such as Facebook advertising, Instagram, WhatsApp, and Messenger, as well as the Meta Business Suite. The Register notes that there are at least one million businesses currently using Meta’s AI agents on WhatsApp and Messenger alone.
The specific clause added to the terms authorizes Meta or its agents to use automated software to access and index content available through domains associated with the customer’s products. This includes periodically pulling updates from these web pages. The company asserts that this data collection is necessary to provide and enhance the features provided under the commercial terms and any applicable supplemental agreements.
Industry Reaction and Impact
The move has drawn immediate condemnation from industry representatives concerned about data privacy and content ownership. Danielle Coffey, president and CEO of the News/Media Alliance, publicly condemned Meta’s decision to expand its data collection rights. The Register highlights that almost every news publisher maintains at least one Meta business account, meaning the majority of digital media outlets fall under the definition of business customers.
Nieman Lab, cited in the reporting, points out that the prevalence of Meta business accounts among publishers makes this policy change particularly significant for the news industry. Publishers using Meta’s advertising or marketing tools may need to reconsider what information they share online, given that Meta will have the contractual right to scrape their associated web properties.
What should you do?
Businesses and publishers using Meta’s commercial services should review the upcoming changes to the terms of service carefully. Organizations should audit their web properties to understand what content is accessible via domains associated with their Meta business accounts. It is advisable to consult legal counsel to assess the implications of granting automated scraping rights to a third party for AI training purposes.
What to do and how to stay safe: Meta
- Review Meta’s updated commercial terms when they are published to understand the specific data collection clauses.
- Audit all domains and web pages linked to your Meta business accounts to identify sensitive or proprietary content.
- Consult with legal experts to evaluate the risks of allowing automated scraping of your associated web properties.
- Monitor communications from Meta regarding the implementation timeline and any opt-out mechanisms that may be offered.
Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
When will these new terms take effect?
Meta plans to introduce the new terms at the end of the month, according to The Register.
Who is considered a business customer under these terms?
Business customers include those using ads, selling products, developing apps, or managing Pages or Groups for business purposes.
Can Meta use the scraped data for AI training?
Yes, the new terms explicitly allow Meta to use collected content to train AI models and improve its products.
