GEO Spam Is Coming. Here's How to Build AI Visibility That Survives It.
What Google, DeepSeek and ChatGPT did about manipulated AI answers in 2026, rechecked on November 19, and what outlasts the next purge.
What Google, DeepSeek and ChatGPT did about manipulated AI answers in 2026, rechecked on November 19, and what outlasts the next purge.

Where the data comes from: the author’s November 19, 2026 search of X, Google’s spam policies, guide to generative AI search and Search Status Dashboard, three articles by Lily Ray, a Fudan University paper on arXiv, DeepSeek’s changelog, and SparkToro’s and Ahrefs’ studies, as the post rechecks them.
- Substack post: GEO Spam Is Coming. Here’s How to Build AI Visibility That Survives It.
- RankControl, an AI SEO and AEO platform
The sources and figures in the post
| Number | What the post says | Source | Date |
|---|---|---|---|
| 20 posts, one account | “I searched X for the phrase "200+ AI SEO prompts" and got 20 posts, all from one account.” | The author’s search of X (accounts not named in the post) | November 19, 2026 |
| 7 promotions | “Seven were promotional posts dated October 2 to October 18, with 1 to 4 likes and 1,621 to 2,684 views each.” | Same search | Posts dated October 2 to 18, 2026 |
| 13 replies | “The other thirteen were word-for-word identical replies” | Same search | Replies including October 1, 2026 |
| Fudan test | “across 10 platforms 8 cited the invented concept within a week” | Search Engine Journal on a Fudan University paper (arXiv) | October 19, 2026 (paper posted October 8) |
| Site-search jump | “jumped from 0.37% to 16.8% in a day” | Suganthan Mohanadasan’s note | August 19, 2026 (jump on August 8) |
| Peec AI jump | “0.3% to 23% on the day GPT-5.6 became the default” | Peec AI tracker, as reported by Lily Ray | August 17, 2026 |
| Reddit share | “Reddit’s share of ChatGPT Search citations averaged 0.52% against 3.83% from July 18 to August 7, an 86% fall” | Promptwatch, as reported by CMSWire | October 21, 2026 |
| Reddit in AI Overviews | “its share in Google’s AI Overviews slipped only from about 2.5% to 2.1%” | CMSWire | October 21, 2026 |
| Listicle study | “Of those, 80 produced an AI Overview, the answers cited 323 self-promotional listicles, and in 224 of them (69%) the publisher was cited as a source but left out of the recommendation” | Lily Ray, 100 B2B “best [category]” queries on April 15, May 15 and June 8, 2026 | June 17, 2026 |
| Listicle update | “found AI Overviews citing 38% fewer of these listicles than in June, leaving the publisher out 83% of the time and helping competitors get recommended in 70% of cases” | Lily Ray, covering September 2026 | October 4, 2026 |
| Policy change | “attempting to manipulate generative AI responses in Google Search” | Google, spam policies, lead paragraph changed May 15, 2026 | Page last updated August 28, 2026 |
| August update | “Google’s Search Status Dashboard shows the August 2026 spam update released on August 18 and complete as of August 21” | Google Search Status Dashboard | August 18 to 21, 2026 |
| llms.txt | “neither harm nor help” | Google, guide to generative AI features | Last updated July 10, 2026 |
| DeepSeek | “filters SEO spam, GEO manipulation, and low-quality AIGC (AI-generated content) through strict detection mechanisms” | DeepSeek’s changelog | October 17, 2026 |
The five phrases worth questioning in a GEO pitch, and the figure behind each
| Phrase | What the post says | Source | Date |
|---|---|---|---|
| “Guaranteed AI citations” | “only 54.5% of cited URLs overlap, so 45.5% are new each time” | Ahrefs, 43,000 keywords, Google AI Overviews only | November 11, 2025 |
| “Guaranteed AI citations” | “had 600 volunteers run 12 prompts through ChatGPT, Claude and Google’s AI a combined 2,961 times” | SparkToro, which found under a 1 in 100 chance of the same list of brands twice | January 27, 2026 |
| “We know how the algorithm picks sources” | “its share in Google’s AI Overviews slipped only from about 2.5% to 2.1%” | CMSWire | October 21, 2026 |
| “Our proprietary network of authority sites” | “creating multiple sites with the intent of hiding the scaled nature of the content” | Google’s spam policies, an example of scaled content abuse | Page last updated August 28, 2026 |
| “Look at these results” | “Ahrefs measured Google AI Overviews changing every 2.15 days on average” | Ahrefs | November 11, 2025 |
| “llms.txt optimization package” | “RankControl added it to 12 sites and tracked AI citations for 90 days, and its July 11, 2026 write-up found no meaningful change from the file alone” | RankControl’s own test (not linked in the post) | July 11, 2026 |
The 2026 enforcement dates
| Date | What the post says |
|---|---|
| March 24 | “lists spam updates on March 24, June 24, August 18 and September 24, 2026” |
| May 15 | “Google changed the lead paragraph of its spam policies” |
| August 8 | “on August 8, 2026 the share of those that name a specific site with the site: operator” |
| October 17 | “DeepSeek’s changelog entry dated October 17, 2026” |
- Reference: https://rankcontrol.github.io/images/ss-042-geo-spam-is-coming-info-timeline-wm.jpg
- Reference: https://thesearchshift.substack.com/p/geo-spam-is-coming-heres-how-to-build
- Reference: https://rctrl.com
- Reference: https://developers.google.com/search/docs/essentials/spam-policies
- Reference: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
Write a comment