AI-scraping "policy token" standard
Where it stands: OpenAI, Google, Microsoft and Anthropic have committed to adopting a shared "policy token" signal so publishers can set AI-scraping preferences across all the companies at once. Anthropic standards lead Kevin Kelley proposed it through the Internet Engineering Task Force, but it is not yet implemented and no technical specification or enforcement plan has been given. Watch the proposal's progress, publisher adoption at scale, whether the commitments reach products, and holdout behavior from other AI companies.
Latest
- 2026-10-31 (The Verge — theverge.com/tech): OpenAI, Google, Microsoft and Anthropic commit to adopting a shared "policy token" signal so publishers can set their AI-scraping preferences across all of the companies at once instead of configuring each lab separately; proposed by Kevin Kelley, a technical lead for standards at Anthropic, as part of the Internet Engineering Task Force process. Announced, not yet implemented — no technical specification or enforcement story in the reporting yet. Watch: the IETF proposal's progress, publisher adoption at scale, whether the commitments land in product, and holdout behavior from other AI companies.