Updated on 14 November 2026, by João Leite, Founder of murmur.
In Google Search, content made with AI is not penalized when the page helps the reader, and what becomes spam is manipulation. In a February 2023 post, Google writes that “O uso apropriado de IA ou automação não vai contra nossas diretrizes.” (“The appropriate use of AI or automation is not against our guidelines.”) The OpenAI and Anthropic pages for site owners deal with the crawlers that read the site, not with how the text was made. Each company talks about its own search, and one company’s rule does not apply to the other.
- Google Search accepts the appropriate use of AI. What violates Google’s spam policies is using automation, including AI, with the main goal of manipulating ranking in the results.
- Google asks for checking and transparency. The Google page updated on 6 October 2026 asks for manual fact-checking and suggests telling the audience how the content was created.
- In the online store, AI-generated product data carries a label. In Google Merchant Center, product titles and descriptions generated with AI must be provided separately and labeled as “gerado com IA” (“generated with AI”).
- OpenAI and Anthropic ask that the search robot be able to reach the page. OpenAI writes that any public website can appear in ChatGPT search, and Anthropic warns that blocking Claude-SearchBot can reduce the site’s visibility in Claude.
- murmur measures whether the page becomes a source. murmur, a GEO agency in Brazil that places brands in the answers of ChatGPT, Claude, Google AIO and other AI agents, tracks this on the same questions, week after week.
Google does not penalize content made with AI when the page helps the reader
Google does not penalize content made with AI when the page helps the reader, and it punishes automation made to manipulate ranking. The Google Search Central post signed by Danny Sullivan and Chris Nelson, from 8 February 2023, carries both halves of the rule. The first is the sentence that the appropriate use of AI or automation is not against the guidelines. The second says that “O uso da automação, incluindo IA, para gerar conteúdo com o objetivo principal de manipular a classificação nos resultados da pesquisa é uma violação das nossas políticas de spam” (“The use of automation, including AI, to generate content with the primary purpose of manipulating ranking in search results is a violation of our spam policies”).
For Google Search, the tool gives the page no advantage. The same post sums it up like this: “Usar a IA não dá vantagens especiais para o conteúdo. É apenas conteúdo.” (“Using AI doesn’t give content any special gains. It’s just content.”) A page written with the help of an AI goes through the same Google Search rules as a page written by hand.
The case of many pages generated to manipulate search has a name in Google’s spam policies, scaled content abuse, and a section of its own in the guide on how many pages to publish for your company to appear in ChatGPT. Google’s sentence about content made with AI applies to its own search, as the guide on what Google says and the reach of its rule shows.
What Google asks of those who publish content made with AI: check the facts and tell how it was made
Google asks two things of those who publish content made with AI: check the facts by hand and, when it makes sense, tell the audience how the content was created. The page “Orientações da Pesquisa Google sobre como usar conteúdo de IA generativa no seu site” (“Google Search’s guidance on how to use generative AI content on your site”), updated on 6 October 2026, writes that “É fundamental verificar manualmente os fatos e revisar todo o conteúdo gerado com IA” (“It’s critical to verify the facts manually and review all AI-generated content”). The same page suggests that publishers consider adding information about how the content was created, in a way that makes sense for their audience.
Google’s requests for content made with AI, putting together the 2023 post and the 2026 page, fit into four points:
- Verify the facts by hand and review all AI-generated content before publishing.
- Tell how the content was created, when readers might ask themselves “How was this created?”.
- Label the product title and description made with AI as “gerado com IA” (“generated with AI”), in Google Merchant Center.
- Do not use automation, including AI, with the main goal of manipulating ranking in search results.
The notice about how the text was made is a suggestion, not an obligation, for site content. The 2023 post says that AI or automation disclosures are useful for content where someone might think “How was this created?”. The 2026 page repeats the idea with the verb “consider”. The labeling obligation appears in one place only, the store’s product data.
Content made with AI in an online store: the label Google Merchant Center asks for
In the online store, content made with AI has one extra rule: product titles and descriptions generated with AI carry the label “generated with AI” in Google Merchant Center. Merchant Center is Google’s dashboard where the store submits its product catalog. The page of 6 October 2026 says that AI-generated product data, such as title and description attributes, must be specified separately and labeled as “generated with AI”.
The difference between the store and the rest of the site is the verb. For site text, Google suggests telling how it was made. For product data, Google requires separating and labeling. A store that writes the description with AI complies with both: it checks the product facts and labels the data in Merchant Center.
What OpenAI and Anthropic write about content made with AI
OpenAI and Anthropic, on their pages for site owners, write about the bots that read the site, not about content made with AI. OpenAI, which makes ChatGPT, writes in its help center that “Any public website can appear in ChatGPT search.” The same page asks that the site not block OAI-SearchBot, ChatGPT’s search bot, so that the content is included in the summaries and excerpts of the answer.
Anthropic, which makes Claude, warns in the Claude help center that blocking Claude-SearchBot prevents the system from indexing the site’s content for search and can reduce the site’s visibility and accuracy in answers. What both companies ask is that the search bot reach the site.
| company | what the page for site owners asks | about which search |
|---|---|---|
| appropriate use of AI is accepted; automation to manipulate ranking is spam; check the facts; tell how it was made; label in Merchant Center | Google Search | |
| OpenAI | any public website can appear; do not block OAI-SearchBot | ChatGPT search |
| Anthropic | blocking Claude-SearchBot can reduce the site’s visibility and accuracy | Claude search |
| Microsoft | content made to manipulate Bing and Copilot language models can lose visibility, according to Bing’s webmaster guidelines (more in the guide on manipulating the answer) | Bing and Copilot |
The rule about content made with AI belongs to Google alone. OpenAI and Anthropic deal with bot access, and Microsoft deals with content made to manipulate the models. Google’s rule, therefore, applies to Google Search, and does not become a rule of ChatGPT, Claude or Perplexity. robots.txt, where the site records what each bot may read, has its own explanation in the guide on llms.txt and what works to be cited by AI.
Content made with AI and the answers from AI models: what counts is the page answering the question
For content made with AI to appear in AI answers, what counts is the page answering the searcher’s question, and the search bot being able to reach it. OpenAI and Anthropic ask for bot access, and Google asks for a useful page, with the facts checked.
The page that serves AI models answers the question right at the start, includes the facts with a number and a source, and covers one subject at a time. How to write that text is in the guide on how to write a text with the answer at the start and numbers with sources. Using ChatGPT to write, on the other hand, is not the same as preparing the site for AI models, as the guide on SEO for AI, the optimization of sites to appear in ChatGPT explains.
murmur measures whether the page becomes a source, on the same questions, week after week
murmur measures whether the page becomes a source for AI answers, not how it was written. Every week, the same questions go back to ChatGPT, Claude, Perplexity and Google AIO, and each round shows whether the page made it into the answer. The result sits in a dashboard included in the service, not sold separately, with the brand’s presence alongside that of each competitor.
murmur treats AI answers as the test of each page. What it produces serves any AI that reads the open web, and the weekly round points out the answers each page made it into. From the diagnostic onward, one person looks after only that brand: they know the company and the sector, read the results and follow every change closely.
Frequently asked questions
Can Google tell that the text was written by ChatGPT?
Google’s search rules do not depend on knowing which tool wrote the text, but on the purpose of the page. Google’s post of 8 February 2023 says that appropriate use of AI or automation does not go against the guidelines and that using automation, including AI, with the primary goal of manipulating ranking violates the spam policies. The same post says that using AI gives no special advantage to the content, which is just content.
Do I need to disclose that the text was made with AI?
In the text on the site, the disclosure is a Google suggestion, not an obligation. Google’s page of 6 October 2026 suggests that publishers consider adding information about how the content was created, in a way that makes sense to the audience. The labeling obligation applies to store product data in Google Merchant Center, marked as “gerado com IA” (generated with AI).
Can I use AI to write the store’s product descriptions?
The product description can be written with AI, and Google Merchant Center asks that the AI-generated title and description come separately and labeled as “gerado com IA” (generated with AI). Google’s page of 6 October 2026 also asks that facts be checked by hand and that all AI-generated content be reviewed. In a store, that means checking measurements, price and features before publishing.
Does a page made with AI appear in Google AIO?
A page made with AI can appear in Google AIO, the AI-written summary that Google shows at the top of many searches, before the links. To become a supporting link in the summary or in AI Mode, Google’s conversational search, the page must be indexed and eligible to appear in Google Search with a snippet, according to Google’s page of 31 December 2025. The 2023 post says that using AI gives no special advantage to the content, which is just content, and so a page made with AI goes through the same conditions.
Do ChatGPT and Claude penalize content made with AI?
The OpenAI and Anthropic pages for site owners do not talk about content made with AI, but about the search bot. OpenAI writes that any public website can appear in ChatGPT search and asks that OAI-SearchBot not be blocked. Anthropic warns that blocking Claude-SearchBot can reduce the site’s visibility and accuracy in answers with search. Google’s rule on content made with AI applies to Google Search.
Where to start
Start by seeing whether your company’s pages already make it into AI answers. murmur’s free diagnostic asks only for the site address, with no sign-up, and shows in less than two minutes how AI models answer the questions of people looking for what your company sells; the full report arrives by email. murmur’s sector studies show which brands AI models recommended in the sectors murmur measured.
Read also
- How many pages to publish so your company shows up in ChatGPT: one per unanswered question
- Manipulating what ChatGPT recommends: it worked in a test and puts your site at risk
- How to write text to appear in ChatGPT: answer up front and numbers with sources
- Optimizing for AI Is Not Just SEO: Google Only Talks About Its Own Search
References
- Google, Danny Sullivan and Chris Nelson, “Orientações da Pesquisa Google sobre conteúdo gerado com IA”, Google Search Central blog, 8 February 2023. “O uso apropriado de IA ou automação não vai contra nossas diretrizes.” (“Appropriate use of AI or automation does not go against our guidelines.”); automation to manipulate ranking violates the spam policies; disclosure is useful when someone might think “Como isso foi criado?” (“How was this created?”). https://developers.google.com/search/blog/2023/02/google-search-and-ai-content?hl=pt-br (accessed on 7 November 2026).
- Google, “Orientações da Pesquisa Google sobre como usar conteúdo de IA generativa no seu site”, Google Search Central, updated on 6 October 2026. “É fundamental verificar manualmente os fatos e revisar todo o conteúdo gerado com IA” (“It is essential to manually verify the facts and review all AI-generated content”); state how it was created; “generated with AI” label on product data. https://developers.google.com/search/docs/fundamentals/using-gen-ai-content?hl=pt-br (accessed on 7 November 2026).
- Google, “Políticas de spam para a Pesquisa Google na Web”, Google Search Central, updated on 16 September 2026. Scaled content abuse: many pages generated with the primary goal of manipulating search rankings rather than helping users. https://developers.google.com/search/docs/essentials/spam-policies?hl=pt-br (accessed on 1 October 2026).
- Google, “Recursos de IA e seu site”, Google Search Central, updated on 31 December 2025. The page must be indexed and eligible to appear in Google Search with a snippet to become a supporting link. https://developers.google.com/search/docs/appearance/ai-features?hl=pt-br (accessed on 1 October 2026).
- OpenAI, “Publishers and Developers - FAQ”, OpenAI help center, undated page. “Any public website can appear in ChatGPT search.”; do not block OAI-SearchBot. https://help.openai.com/en/articles/12627856-publishers-and-developers-faq (accessed on 30 September 2026).
- Anthropic, “Does Anthropic crawl data from the web, and how can site owners block the crawler?”, Claude help center, undated page. Blocking Claude-SearchBot can reduce the site’s visibility and accuracy in answers with search. https://support.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler (accessed on 30 September 2026).
- Microsoft, “Bing Webmaster Guidelines”, Bing Webmaster Tools documentation, undated page. Trying to manipulate the language models used by Bing or Copilot can reduce visibility or remove the site from search. https://www.bing.com/webmasters/help/webmaster-guidelines-30fba23a (accessed on 1 October 2026).