Google has updated its public advice on content written with AI tools. In early October it said that all AI-generated content should be checked by a person before it is published, and that invented author profiles count as deception. If your business publishes blog posts, product pages or service pages with help from AI, these updates may already apply to your site. This article explains what changed, how Google judges content now, and what you can do about it.

What Google changed in October

Google publishes its advice on Search Central, its official help site for website owners. Two of its guidance pages changed in the first week of October.

  • Fact-checking is now expected. Google’s guidance on using generative AI content, updated October 1, says it is “critical to manually factcheck and review all AI-generated content.” Generative AI means tools such as ChatGPT that write new text from a short instruction.
  • The check covers the small text too. The review also applies to metadata, the background text such as a page’s title, its search result description and its image descriptions (called alt text).
  • Fake authors are named as deception. Google’s page on helpful, reliable content now warns against “AI-generated headshots, made-up names, or false credentials.” Search Engine Journal, a trade publication for search marketers, reported the addition on October 2.
  • “Main content” now has a definition. The same page describes main content as the part of a page that directly helps it do its job, and judges it on effort, originality, skill and accuracy.

How Google judges AI-generated content

Google does not ban AI writing. Its guidance says generative AI “can be particularly useful when researching a topic,” and for adding structure to original work. What Google looks at is the finished page, whoever or whatever wrote it.

The four qualities it now names for main content work like this:

  1. Effort. How much human work went into the page. Google says mass-producing text with AI and no human oversight counts as little or no effort.
  2. Originality. Whether the page offers information or a point of view that isn’t available elsewhere.
  3. Skill. Whether the page shows the knowledge needed to answer the question well.
  4. Accuracy. Whether the facts are right. Google holds pages about health, money and safety to a higher standard.

For example, a plumbing company asks an AI tool for a post on how to keep pipes from freezing. The draft reads well, but it gives the wrong freezing temperature and recommends a product the company doesn’t sell. Nobody checks it before it goes live. Under the new guidance, that page shows low effort and low accuracy, even though the writing is smooth.

Why this matters even if you don’t use AI yourself

Many businesses get content from a freelancer, an agency or a website plugin, and any of them may use AI tools without saying so. Google’s guidance is about the page on your site, so the same standards apply no matter who wrote it.

Author information is the other common gap. Some website themes and content services add a generic author, such as “Admin” or a first name with a stock photo. Google’s page asks whether pages carry “a byline, where one might be expected,” meaning a line that names the person who wrote it. A missing byline is not the same as a fake one. An invented person with an AI-made photo, though, is now named directly as deceptive.

Accuracy also matters for the AI-written summaries Google shows for many searches, which we explained in our post on AI Overviews for brand name searches.

What is known so far

  • Google is running more spam updates. A spam update is a change to Google’s systems that finds and demotes low-quality pages. Google’s Search Status Dashboard lists four in 2026, the most recent running from September 24 to about October 7.
  • Mass-produced pages are a stated target. The generative AI page says using AI to “generate many pages without adding value for users” may break Google’s spam policy. Google calls this scaled content abuse.
  • Disclosure is optional. Google suggests that sites using automation “consider adding information on how your content was created.” It doesn’t require a label.
  • Some things are still unclear. Google has not said how it spots fake author profiles. For now, the guidance describes what Google wants, not how it checks.

Your three options for AI-written content

  1. Keep using AI, with a human review step. A named person checks facts, claims and metadata before anything is published. This keeps most of the time savings, and it adds editing time to every page.
  2. Use AI only for research and outlines. People write the final text. This matches the uses Google describes as helpful, and it costs more writer time.
  3. Stop using AI for published content. This removes the question, and it gives up the time savings.

Any of these can be the right choice for a given business. What all three share is that a real person is responsible for each page. We covered how to keep a consistent voice when AI does some of the writing in our post on brand voice guidelines and AI.

Where to start

Most of this is a review of what is already on your site, and a week or two is usually enough for a small one. A reasonable first pass looks like this:

  • List your AI-assisted pages. Ask your team and any vendors which pages were written with AI tools in the last year.
  • Fact-check the pages with the most visitors. Start with service pages and popular posts, and include their titles, descriptions and image alt text.
  • Check every author profile. Each byline should name a real person, with a real photo and an accurate job title.
  • Write down a review rule. Decide who checks AI-assisted content before it goes live, and what they check.

If you’d like help with any of it, that’s work we do every day.