Meta postpones new publisher terms amid AI training concerns
Meta has postponed new business terms after publishers and other industry groups warned they could let it scrape news websites and use their material to train AI. The delay gives the company more time to explain the changes and address concerns about how website content is accessed.
The terms were due to take effect on 30 October and are now scheduled for 20 November. Meta removed wording that referred to training its products, but legal experts said this did not rule out the possibility; the News/Media Alliance called for revisions, while groups in the UK and EU urged regulators to intervene. Meta says site owners can still use robots.txt to signal whether crawlers may access their content.
In short
- Meta delayed the terms until 20 November.
- Publishers fear the terms could permit AI training on their content.
- Meta says websites can signal crawler preferences with robots.txt.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Meta could argue that clear, consistent terms and crawler access help it develop AI products while giving publishers a way to control access through robots.txt. Postponing the change and removing explicit training language also give the company time to explain its approach and respond to concerns.
The case against
Publishers could argue that removing the training reference does not settle whether their material may still be used to train AI, leaving uncertainty about consent, compensation and control. They may also contend that robots.txt is not an adequate safeguard if its effect is unclear or difficult to enforce, and that regulators should ensure the terms respect publishers’ rights.
AI Business Companies Technology
Read the full article at the source →
Originally published by The Verge as “Meta delays new business terms after publisher pushback on AI scraping”.