TL;DR: Meta has revised its business-account terms to cover the use of web scrapers (programs that extract data from websites), bots, and crawlers to access, analyze, and use data to develop AI, while questions remain about the scope of robots.txt and publishers’ rights.
If I were to put it simply, I’d say this is not merely a wording change in the advertising terms. It expands the area of data that Meta may access from businesses, especially news organizations. It is like giving the big player permission to inspect a store’s data warehouse, even while claiming to respect the “no entry” sign.
Meta has revised the terms for business accounts using Facebook, Instagram, WhatsApp, and Messenger, allowing web scrapers, bots, and crawlers to access related websites, as well as index, analyze, and extract content to develop, research, and train AI. The new terms are scheduled to take effect on October 30.
Meta says it has not changed how it uses web crawlers and will continue to support robots.txt as before. On October 17, the company clarified that the terms apply to advertising and business products, while revising the wording to define the scope more clearly.
New terms raise questions across the news industry
The issue is not limited to crawling websites. It also includes creating test credentials to access, index, analyze, store, and extract data from domains associated with business accounts, including periodically retrieving updated data.
Legal experts believe these terms could allow Meta to claim authority over access restrictions such as robots.txt and use news content to compete directly with news organizations. At the same time, businesses that do not accept the terms could have their access to Meta’s business products suspended or restricted.
When news websites could be used to train AI without realizing it
Imagine a news organization using a business account on Facebook, Instagram, WhatsApp, or Messenger, then discovering that the new terms may also cover its own website. Meta states that it may use automated software to crawl and collect content for developing, researching, and training AI.
The key question is how much authority robots.txt still carries if the business terms are interpreted as taking precedence over existing blocking instructions, and whether Meta could use the content without entering into a licensing agreement. The issue becomes even more serious when that content could be used to compete with the news organization that originally published it.
From advertising terms to access rights for business websites
These terms fall under the Commercial Terms for business accounts that promote content through Facebook, Instagram, WhatsApp, and Messenger. They are therefore not limited to purchasing advertisements on the platforms.
Meta explains that it is “consolidating” the commercial advertising terms and self-serve advertising terms into a single set. The reason is to make its business requirements cover the use of Meta’s products and services more broadly.
How do the previous and new terms differ?
The article reports that the new terms may allow Meta to crawl and use content from websites associated with business accounts to develop and train AI, and may also allow it to challenge robots.txt instructions. Meta, meanwhile, maintains that it has not changed how it uses web crawlers and will continue to respect robots.txt.
| Factor | Article's report | Meta's clarification |
|---|---|---|
| Crawling scope | Websites related to the business, including news websites | Focuses on product information for advertising and recommendations on the platform |
| robots.txt | May challenge blocking instructions | Website owners can still use robots.txt to state their preferences |
| Wording revision on October 17 | Questions remain about the scope of use | Clarifies that the crawling terms apply to ads and business products |
How can a single phrase affect news publishers?
The terms give Meta the authority to use bots and crawlers to access domains or webpages associated with business accounts, such as a news organization’s website that promotes content on Facebook, Instagram, WhatsApp, and Messenger.
Meta may also periodically retrieve data from these websites to develop, research, train, and maintain AI. If the content is displayed or used in a way that competes with the news organization, publishers may have to deal with Meta services that use data from their own websites, without a licensing agreement covering all of the content.
Meta and OpenAI’s approach to copyright agreements
| Factor | Meta | OpenAI |
|---|---|---|
| Apparent approach | Relies on commercial terms to crawl websites associated with business accounts | Enters into AI licensing agreements |
| Named news companies | No licensing partners are identified in this section | CNN, Fox News, USA Today, and People Inc. |
| Deal terms | The article describes the crawling terms but provides no licensing details | The article does not provide details about the terms of each deal |
Overall, Meta is being questioned over its use of commercial terms to access content, while OpenAI uses licensing agreements with named publishers. However, the article does not explain how much content OpenAI’s deals cover or what usage conditions apply.
Strengths and remaining ambiguities of this policy
One strength is that Meta says the policy can help retrieve product information for more accurate advertising and recommendations on its platforms. However, the scope of the phrase “related” websites associated with a business account is very broad, raising questions about how much news content could be accessed.
Pros
- +Helps retrieve product information for more accurate advertising and recommendations
- +Meta says it continues to support the use of robots.txt to express content owners’ preferences
Cons
- −The scope of accessible websites and content remains broad and unclear
- −May allow content to be used to train AI without a specific agreement
- −There is still debate over whether the commercial terms take precedence over robots.txt
Costs that do not appear on the invoice
The most significant cost may not be a service fee, but losing control over how content will be used by others.