Are Automated Blog Posts Effective for SEO? What Google Says and What Works

Google's own summary is blunt. "Using AI doesn't give content any special gains. It's just content," its Search team wrote in its guidance on AI-generated content. Large studies fit that picture: AI-heavy pages do rank, but on average they rank lower and earn fewer impressions than pages people wrote or heavily edited.

A disclosure first: Rankbox sells an AI article writer, so weigh this with that in mind. Everything below is sourced to Google or to independent studies, and the product gets one paragraph near the end. For the bigger picture, including why generic AI articles are so easy to replace, read our deeper piece on synthetic content saturation and the model collapse moat.

What Counts as Automated Blog Posts?

People mean very different things by "automated blog posts." Google treats them differently too, so it helps to name which one you're asking about.

So the honest answer depends on the row. Scheduling is neutral. Drafting with AI is fine if a person does the work of checking and adding. The trouble starts when the draft is the final product and there are hundreds of them.

What Google Says About Automated Blog Posts

Google's position hasn't changed much since 2023. What has changed is the spam policy's wording, which since March 2024 targets content made at scale however it's produced.

"It's just content": the 2023 guidance

Google's February 2023 post says "Appropriate use of AI or automation is not against our guidelines." It adds that AI content "might do well in Search" if it is "useful, helpful, original, and satisfies aspects of E-E-A-T. If it doesn't, it might not." E-E-A-T stands for experience, expertise, authoritativeness and trust; see our E-E-A-T entry.

The same post gives a one-line test for intent: "If you see AI as an inexpensive, easy way to game search engine rankings, then no." On labels, Google says AI disclosures are "useful for content where someone might think 'How was this created?'" They're encouraged, not required.

The scaled content abuse policy, as of 28 September 2026

Google's spam policies page was last updated on 28 August 2026. It defines scaled content abuse as "when many pages are generated for the primary purpose of manipulating search rankings and not helping users." It is "typically focused on creating large amounts of unoriginal content that provides little to no value to users, no matter how it's created." The listed examples are:

1. Using generative AI or similar tools to produce many pages that add no value for users.
2. Scraping feeds, search results or other content to make many pages, including by synonymizing, translating or other obfuscation.
3. Stitching together content from different pages without adding value.
4. Creating multiple sites to hide how much content is being scaled.
5. Creating many pages that make little sense to a reader but contain search keywords.

Two details are easy to miss. The page now defines spam to include "attempting to manipulate generative AI responses in Google Search," which reaches answers in features such as AI Overviews and AI Mode. And it says Google detects violations "through automated systems and, as needed, human review that can result in a manual action." Violating sites "may rank lower in results or not appear in results at all." Our glossary covers scaled content abuse in more depth.

The questions Google asks about automation

Google's people-first content page lists warning signs. Two bear on automation directly: "Are you using extensive automation to produce content on many topics?" and "Are you producing lots of content on many different topics in hopes that some of it might perform well in search results?"

Its generative AI page also points to the Search Quality Rater Guidelines, section 4.6.5 on scaled content abuse and section 4.6.6 on content with "little to no effort, little to no originality, and little to no added value." Google says rater scores don't directly change rankings, but the guidelines show what its systems are built to recognize.

Spam updates keep coming in 2026

Google's Search Status Dashboard lists spam updates on 24 March, 24 June and 18 August 2026. A fourth, the September 2026 spam update, was released on 24 September and "applies globally and to all languages." As of 28 September 2026 it's still rolling out, which can take up to two weeks. If you publish automated blog posts at volume, watch Search Console closely until it finishes.

What the Evidence Shows About Automated Blog Posts

Google's rules tell you where the line is. Studies tell you how automated blog posts and other AI-heavy pages perform on each side of it. All of them use AI detectors, which make mistakes, and all of them show correlation, not cause.

Ahrefs: no penalty, but a clear slope

In 2025, Ahrefs checked 600,000 pages ranking in the top 20 for 100,000 random keywords. Only 13.5% scored as purely human-written, 4.6% as purely AI and 81.9% as a mix. The correlation between AI share and ranking position was 0.011, which is effectively none.

A follow-up in July 2026 looked closer and found a gentle slope. Some fully AI pages reach the top: 5.3% of pages in positions one to three scored as 100% AI. But performance fell as the AI share rose:

• Indexing: 49.28% of low-AI pages were indexed, against 40.35% of very-high-AI pages.
• Impressions: low- and moderate-AI pages earned 2–3 times the impressions of high-AI pages.
• Top spots: pages under 50% AI content held 82.2% of top-three rankings.

Ahrefs saw no sign of a switch that blocks AI content. Its author's reading is that AI use tends to go with lower quality, and quality is what Google measures.

Graphite: people-written pages hold the top spots

Graphite's October 2025 study of Google results for 31,493 keywords found 86% of ranking articles were human-written. In ChatGPT and Perplexity, 82% of cited articles were human-written. Only 7% of the articles in position one were AI-generated. Graphite didn't test AI drafts that people heavily edited, and says that approach "may be an effective strategy."

Newer data is less tidy. For a set of questions mostly about named products and people, Graphite found that 42.7% of ChatGPT's cited pages in June 2026 were predicted to be AI-generated. Different questions and detectors give different shares.

What a scaled-content crash looks like

Consultant Glenn Gabe calls the pattern "Mt. AI": traffic climbs fast on scaled AI content, then falls hard. In an April 2026 case study, he described a site whose local news directory had over 850,000 indexed URLs of fully AI-generated articles. After a manual action for scaled content abuse, the section dropped out of Google, and its visibility in AI Overviews, AI Mode and ChatGPT citations fell with it. Engines that ground answers in search results can lose you when search does.

The Keep-or-Cut Test for Automated Blog Posts

The Keep-or-Cut Test is our eight-question check for any post that a tool drafted or scheduled. Answer yes or no for each. It takes about ten minutes a post once you're used to it.

1. Real question. Would a customer actually ask this? Check Search Console, sales calls or support tickets.
2. Something new. Does the post say at least one thing the top five results don't? See information gain.
3. Your own fact. Does it include a fact only your company can supply, such as a price, a spec or a test result?
4. Checked by a person. Has someone who knows the subject verified every claim, number and link?
5. No twin. Is this the only page on your site that answers this question?
6. Sources inline. Does every statistic link to where it came from?
7. Clear owner. Is it clear who stands behind it, with a disclosure where a reader would ask "How was this created?"
8. Pace you can review. Are you publishing no faster than your team can check?

Score automated blog posts the same way every time: 7–8 yes means publish. 5–6 means fix the gaps first. 4 or fewer means cut it, merge it into a stronger page, or keep it out of search.

Worked example: Tallyfold's first 30 drafts

Tallyfold is a made-up invoicing and payments app for agencies. Its team sets up a tool to draft 30 automated blog posts a month and runs the Keep-or-Cut Test on the first batch.

The totals: 12 + 9 = 21 posts published, and 7 + 2 = 9 cut or merged, for 30 drafts. The 21 automated blog posts that shipped each carry at least one fact only Tallyfold holds. The most common failure wasn't bad writing. It was check 3, failed by every draft in the fix pile: a draft with no first-party fact reads fine and adds nothing. Tallyfold and its numbers are illustrative.

Where Rankbox Stands on Automated Blog Posts

Here's the one paragraph on our own product. Rankbox's Citation-Ready Writer researches the live web and writes 2,000–3,500-word, source-backed articles, and its Brand Voice feature applies the tone, audience, style rules and product details you give it, mentioning your product where it fits. Articles reach your site through Rankbox's API on the Business plan, $49.50 a month with a 7-day trial (see pricing). It doesn't import your proprietary data, run benchmarks, track AI citations or publish straight to a CMS, so checks 3, 4 and 8 of the test are yours to do, and the test applies to its drafts exactly as it does to any other tool's.