Meta has removed wording from its new business terms that appeared to let it train AI models on content from websites linked to business accounts, but legal experts say the change may not go far enough, The Verge reports.
The terms take effect on 30 October. After objections from publishers, Meta dropped the word "train", yet the revised version still allows it to use such content to develop "business products". Experts quoted by Nieman Lab, including Cornell law professor James Grimmelmann and attorney Matt Topic, argue that this wording is broad enough to cover AI training and could be used to justify scraping whole sites.
A Meta spokesperson told The Verge that site owners can still use robots.txt to signal their preferences, and said the update is not about scraping news sites and does not override those settings.
Source: The Verge
