AI giants back one shared signal to opt out of training
OpenAI, Google, Microsoft and Anthropic plan to honor one shared signal that lets websites opt out of AI training, The Verge and SiliconANGLE report. Today, sites must set that choice for each company. The signal would sit in a site's robots.txt file.
OpenAI, Google, Microsoft and Anthropic say they will support a new shared signal that lets websites opt out of having their content used to train AI, The Verge and SiliconANGLE report. Publishers could set the choice once for many AI companies.
Right now, sites must signal their AI training choices separately to each company, The Verge reports. SiliconANGLE says a site that wants to block training scrapers has to edit its robots.txt file many times, once for each scraper.
The new "policy token" would live in that robots.txt file, which has long signaled to crawlers what they may visit. Kevin Kelley, a standards lead at Anthropic, proposed it through the Internet Engineering Task Force, The Verge reports. SiliconANGLE says a draft has been sent to the group, which maintains the format.
The companies say their AI crawlers will hold back from a site when they find the token, The Verge reports. Anthropic has already adopted it.
Microsoft says it will add finer controls, including a setting called AI grounding, the outlets report. Grounding means pointing an AI tool to a trusted source to cut down on mistakes. Sites will still appear in Bing search results whatever they choose, Microsoft says, according to The Verge.
Meta communications head Andy Stone signaled support but did not say whether Meta would honor the signal. "We believe creating a common signal that conveys such preferences to multiple providers at once is an important step," Stone said, The Verge reports.
What we know: Four major AI companies plan to honor a single opt-out signal for AI training.
What we don't: Exactly when each company will start, and whether Meta will join.
Where outlets agree The Verge and SiliconANGLE report OpenAI, Google, Microsoft and Anthropic will support one shared AI training opt-out signal.
Where they differ The Verge says Microsoft will add support next year with Google, while SiliconANGLE says Microsoft will do so in the coming months alongside OpenAI.