Meta’s wording of its terms governing automated website browsing has raised concerns that content from business websites could be used to train its artificial intelligence models.
اضافة اعلان
In the updated version, the company removed the word “training” and limited the stated use of collected material to its advertising and commercial products.
The change comes as publishers grow more cautious about how AI companies use their work. An analysis by Axios of 102 major US publishers found that 42% block crawlers designed to train AI, compared with 25% that block crawlers used for AI-powered search.
Why did the wording raise concerns?
Meta’s previous terms allowed automated programs to collect content from websites linked to business accounts. The company stated that the material could be used to “provide, develop, improve, research, train and maintain the safety” of Meta products.
The new wording removes “train” and states that the content may be used to “provide, develop, improve, research and maintain” its advertising and commercial products.
Legal experts who spoke to Nieman Lab expressed concern about the broad scope of the previous wording and the possibility that it could have authorized Meta to use publishers’ website content to train AI systems.
Meta rejected that interpretation, saying the terms were intended for its advertising and commercial products and were not designed to create a new way to collect news website content for AI training.
Meta can still crawl those websites automatically. The company says website owners can set their preferences for content collection through a robots.txt file. However, legal experts point out that the revised terms do not explicitly require Meta to follow these instructions.
Publishers seek greater control over AI training
Meta has faced similar concerns over the use of publicly available content to train AI systems. In some regions, the company uses public posts on Facebook and Instagram for this purpose. Users have also previously reported difficulties when attempting to object to the use of their data.
Publishers are also exploring ways to receive compensation when AI companies use their work. For example, reports have indicated that Amazon considered launching a marketplace that would allow media companies to license their content directly to AI companies, while similar licensing agreements are becoming increasingly common across the industry.
