BR /Future of Work · Web Publishing
Microsoft separates publisher preferences for AI training and grounding
Bing's new controls distinguish model training from generated-answer grounding, with a transition beginning in February 2027.
What happened
Microsoft announced independent web-publisher preferences for AI training and grounding on October 30. Training preferences can apply through robots.txt or individual pages; grounding preferences are page-level controls. Microsoft says these choices do not change eligibility for traditional Bing search results.
The new preferences begin taking effect as Microsoft crawls pages from February 1, 2027. Publishers using the older NoArchive or NoCache controls face a March 1 transition deadline, including additional Bing Webmaster Tools steps if they wish to maintain grounding restrictions.
BELL RIVER perspective · Editorial analysis
What it means for business
Our assessment is that this creates a concrete coordination task for content owners and web teams: decide permitted uses, translate them into settings and verify discovery. Search visibility and permission to inform generated answers now require separate discussions.
- Inventory existing NoArchive and NoCache settings.
- Decide training and grounding preferences separately.
- Assign an owner for the transition.
The shared training syntax remains draft work; settings take effect after crawling, not immediately.
Sources & editorial note
This brief summarises a third-party announcement. The business perspective is BELL RIVER’s interpretation, not a vendor claim or a report of a BELL RIVER client project.