Published 12 hours ago • loading... • Updated 1 hour ago
AI giants promise publishers new way to block some scraping
On Friday, OpenAI, Google, Microsoft, and Anthropic pledged to give websites a single method to prevent their content from being scraped to train AI models, with companies restricting crawlers that encounter the shared token.
Publishers currently manage numerous blocking commands for every individual bot crawling the web, a burden the new protocol eliminates by allowing them to apply AI training preferences across multiple companies simultaneously.
Kevin Kelley, a technical lead at Anthropic, proposed the shared "Policy token." signal for the Internet Engineering Task Force; Meta spokesman Andy Stone conveyed support, encouraging cross-industry adoption.
Claude-Maker Anthropic has "already started" rolling it out, while ChatGPT-maker OpenAI plans implementation "in the coming months," and Microsoft and Google will support the standard beginning Feb 1, 2027.
While companies like Cloudflare already block bots by default, publishers including the New York Times have pursued legal action over unauthorized scraping; Microsoft noted opting out will not prevent websites from appearing in search results.
The AI giants promised on Friday to the media and other websites a unique way to prevent their content from being sucked into training their models, on the back of a judicial arm with a part of the press. ...