# Meta revises AI training terms after publishing backlash

> Meta removed wording that said it could train AI on scraped news content, but experts say the new language still leaves room for such use.

Oossa · 2026-10-22 · https://oossa.com/en/meta-revises-ai-training-terms-after-publishing-backlash

Meta changed its business terms to drop the phrase that it may “train” its AI on any content from sites linked to a business account. The updated wording still lets Meta “use that scraped content to develop any ‘business product,’” which experts say can include AI training. The changes take effect on October 30, 2026. Meta says publishers can still block crawling with robots.txt files.

## The facts

- The new terms become effective on 2026-10-30 – “The terms, which go into effect October 30th, stated…”.
- Legal analysts note the revised language still permits use of scraped content for “develop” purposes.

## Why it matters

Publishers may still have their articles used to improve Meta’s AI products despite the wording change.

## Sources & references

1. [Meta’s AI could still scrape news sites, legal experts say](https://www.theverge.com/tech/1017082/meta-terms-update-ai-training-concerns) – The Verge, 2026-10-22

Last updated: 2026-10-22