OpenAI plans shared training-preference signal, not yet implemented

Announced October 30, with implementation planned in coming months; existing crawler controls remain.

OpenAI plans to honor a shared robots.txt training-preference signal in coming months, allowing sites to express choices to multiple participating providers. The exact effective date is unknown. The related IETF document remains a draft, not a completed standard or a guarantee that all AI providers comply.

Existing GPTBot training and OAI-SearchBot search controls are independent. This announcement guarantees neither search visibility nor deletion of prior training data. No website settings were changed or tested for this report.

https://developers.openai.com/api/docs/bots
https://datatracker.ietf.org/doc/draft-kelley-aipref-training-preference-signal/

Usage conditions

Hong Kong: Website operators in Hong Kong can follow this training-preference announcement. No regional exceptions are stated; implementation dates and local adoption are unknown. This is not ChatGPT web, app or API access. No paid plan is announced; Traditional Chinese tooling is unverified.
Macau: Website operators in Macau can follow this training-preference announcement. No regional exceptions are stated; implementation dates and local adoption are unknown. This is not ChatGPT web, app or API access. No paid plan is announced; Traditional Chinese tooling is unverified.
Taiwan: Website operators in Taiwan can follow this training-preference announcement. No regional exceptions are stated; implementation dates and local adoption are unknown. This is not ChatGPT web, app or API access. No paid plan is announced; Traditional Chinese tooling is unverified.

Official source: Making it easier for websites to express AI training preferences

Source published: 2026-10-30

Source checked: 31 Oct 2026 02:24 (UTC+8)

Published here: 31 Oct 2026 02:26 (UTC+8)