noai
noai is an unofficial tag a website can add to say it doesn't want its content used for AI. No standard backs it, so most AI bots ignore it.
noai is an unofficial tag that a website can add to its pages to say it doesn't want its content used by AI. It's usually written as a robots meta tag, <meta name="robots" content="noai">, or sent as a header. It looks like a rule, but no standard stands behind it, so an AI company only follows it if it chooses to.
That's the part to remember. If you're relying on noai to keep your pages out of AI models, you may not be protected at all.
Where it came from
The art site DeviantArt introduced noai and noimageai in 2022, so artists could say their work shouldn't be used to train image models. Its terms ask anyone training on DeviantArt content to respect the tags. Other sites and plugins copied the idea.
None of the search engines' or AI companies' official lists of supported rules include it. Google's list of robots meta rules, for example, doesn't mention noai.
What to use instead
If you want to keep AI companies from training on your pages, block their training bots in your robots.txt. Each company publishes the names of its bots, and most run separate bots for training and for search.
For example, Fernhill Coffee might be happy for ChatGPT to find and cite its brewing guides, but not want them used to train OpenAI's models. It would block GPTBot, OpenAI's training bot, and leave OAI-SearchBot, its search bot, allowed. A noai tag on the same pages wouldn't change anything that robots.txt doesn't already say.
Be careful not to block too much. Keeping a search bot out can keep you out of that platform's answers.
In AnswerOrbit
The Audit tab in Website reads your robots.txt every day and shows whether each of the 42 AI bots we know is allowed, partly allowed or blocked, along with each bot's purpose. It reads robots.txt only, not the meta tags on your pages.