Does blocking GPTBot remove my brand from ChatGPT?

Not entirely. Models already trained on older crawls still 'know' you, and live-search bots (like OAI-SearchBot) are separate from the training bot (GPTBot). But blocking the crawlers that feed training and live retrieval makes it harder for the engine to learn current facts about you or cite your pages in answers. For most consumer brands that want AI visibility, that's the wrong trade.

How do I control which AI bots can access my site?

Through robots.txt, using each bot's user-agent token (GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Google-Extended, PerplexityBot, and others). You can allow all, block all, or allow some and block others. Per bot, per path.

Should you let AI crawlers index your site?

Every brand's robots.txt quietly answers a question most marketing teams have never been asked: are AI engines allowed to read your website? It's a one-line decision with real consequences for whether ChatGPT, Gemini, Claude, and Perplexity can learn about your brand and cite your pages.

The bots aren't all doing the same job

It helps to think in terms of what a crawler is for. The same company often runs several bots across the AI lifecycle:

Training crawlers: e.g. GPTBot (OpenAI), Google-Extended (Google), ClaudeBot (Anthropic), CCBot (Common Crawl). These gather content that may inform future model training. Blocking them means the next generation of models learns less about you.
Live-retrieval / answer crawlers: e.g. OAI-SearchBot and ChatGPT-User (OpenAI), PerplexityBot (Perplexity). These fetch pages at answer time to ground and cite a response. Blocking them means you're less likely to be cited in live AI answers.
Classic search crawlers: Googlebot, Bingbot. These index for search, and increasingly feed AI surfaces like Google's AI Overviews and AI Mode too. One bot, several jobs.

The nuance: a single token like Googlebot can span training-adjacent, indexing, and AI-answer use: which is why "just block the AI bots" is rarely as clean as it sounds.

The real trade-off

Blocking AI crawlers is sometimes defensible, if you're protecting proprietary content, or you simply don't want your writing in training sets. But for a consumer brand that wants to be discovered, the calculus usually runs the other way:

If engines can't crawl you, they fall back to what others say about you. Retailers, reviews, competitors, and you lose control of the narrative.
If retrieval bots can't fetch your pages, you forfeit citations to whoever they can read.
The cost of allowing them is mostly bandwidth; the cost of blocking them is invisibility.

For most brands, the goal isn't to keep AI out. It's to be the clearest, most-citable source AI can find.

How to set it, concretely

It's all in robots.txt, addressed per user-agent. To welcome everyone (the simplest stance):

User-agent: *
Allow: /

To allow most but block a specific training bot, add a group for it:

User-agent: GPTBot
Disallow: /

User-agent: *
Allow: /

A few rules of thumb:

Decide per bot, not "AI: on/off." You might welcome live-retrieval bots (citations) while making a separate call on training bots.
Don't accidentally block them. Many sites disallow AI crawlers by inheriting an old rule or a security plugin's default, then wonder why they're invisible.
Allowing access is necessary, not sufficient. A bot that can read a page still needs to be able to parse it. Clean, semantic HTML and structured data are what turn access into citations.

What to do next

Open yourdomain.com/robots.txt and read it as an AI engine would: which of the bots above are allowed, and which are silently blocked? If you're a brand that wants to show up in AI answers and you're blocking the crawlers that make that possible, that's the first thing to fix, and the easiest.

Then make sure the pages they can reach are actually built to be understood. That's a bigger topic, and exactly what a site-readiness audit, and ongoing monitoring with Buffy Intel, are for.