Google and OpenAI back one shared switch for websites to block AI training
Website owners would no longer need separate instructions for each AI company to keep their pages out of training data.
Google and OpenAI said Oct. 30 that they plan to honor one shared signal that lets website owners keep their pages out of AI training, in which companies feed huge amounts of web text into AI programs to teach them. Today, sites must write separate instructions for each AI company.
The signal goes in robots.txt, a small file most websites use to tell automated programs, called crawlers, which pages they may visit. The idea comes from a working group at the Internet Engineering Task Force, or IETF, a body that writes technical standards for the internet.
Google said a site that uses the shared signal to opt out will be treated as if it had blocked Google-Extended, Google's own AI training control since 2023. Company-specific controls will keep working, so sites can still pick and choose.
Both companies said the change saves site owners work. Google said publishers, public-interest groups and government regulators had asked for one industry-wide control, and that wide support could turn it into a lasting standard. Google said the plan fits a European Union code under its AI law that treats robots.txt as a main way for sites to state their wishes. OpenAI called it a step toward an open web where owners have simpler ways to choose how their work is used.
The signal is voluntary and only binds companies that agree to it. OpenAI said it will honor the signal in the coming months; Google aims to in 2027.
Source: Google Search Central, "Supporting a shared AI training control for websites," Oct. 30, 2026; OpenAI, "Making it easier for websites to express AI training preferences," Oct. 30, 2026.