---
title: "Black Hat GEO: What China's First GEO Rulebook Bans | Mandato Blog"
description: "A TV show got AI to recommend a fake product, and China wrote GEO rules in response. What black hat GEO is, what the rules ban, and why not to use it."
url: https://mandato.agency/blog/black-hat-geo/
last_updated: 2026-10-25
language: en
---

Blog · GEO Services

# After a TV Scandal, China Wrote the First GEO Rulebook. Do Not Use the Tricks It Bans

By [Shahid Ali](https://www.linkedin.com/in/shahidali-in/) · 24 October 2026 · 7 min read

![Cover image for After a TV Scandal, China Wrote the First GEO Rulebook. Do Not Use the Tricks It Bans](https://mandato.agency/assets/img/blog/black-hat-geo/cover.jpg)

Black hat GEO is any trick that fakes content, reviews, sources or traffic so that AI assistants recommend a brand they would not pick on the facts. In September 2026, a Chinese industry body published what [Search Engine Journal calls the first GEO rulebook](https://www.searchenginejournal.com/china-wrote-the-first-geo-rulebook-after-a-tv-scandal-western-seos-should-take-notes/593199/), and it bans these tricks by name. It came six months after a TV show caught AI models recommending a product that did not exist.

This post covers what happened, what the rules ban, and why you should not use black hat GEO anywhere, even where no rulebook exists. The short answer: it already breaks Google's rules and US rules on fake reviews, it can be exposed in public, and the results do not last.

## What black hat GEO means

GEO, or generative engine optimization, is the work of getting AI assistants such as ChatGPT, Gemini, Copilot and Google's AI Overviews to understand, mention and cite a brand. The honest version does this with true, useful pages and real coverage, much like [answer engine optimization](https://mandato.agency/generative-engine-optimization/answer-engine-optimization/), which shapes pages around the questions buyers ask. Black hat GEO tries to fool the model instead.

The tricks fall into groups:

- **Fake content at scale:** many invented articles, reviews and "expert" pages about a product, posted where AI systems will find them.
- **Corpus poisoning:** flooding the web with false claims so they end up in what a model reads or retrieves.
- **Prompt injection:** hidden instructions in a page that try to tell an AI system what to say.
- **Fake signals:** bought traffic, fake reviews, forged sources and made-up experts.
- **Attacks on rivals:** false negative content about competitors.

These are the tactics the Chinese rulebook lists, according to Search Engine Journal's report of 21 October 2026. Most of them are old black hat SEO tricks pointed at a new target.

![Search Engine Journal's October 2026 article headline saying China wrote the first GEO rulebook after a TV scandal and that Western SEOs should take notes.](https://mandato.agency/assets/img/blog/black-hat-geo/sej-geo-rulebook.jpg)

Search Engine Journal's report on China's GEO rulebook. Screenshot: [Search Engine Journal](https://www.searchenginejournal.com/china-wrote-the-first-geo-rulebook-after-a-tv-scandal-western-seos-should-take-notes/593199/), taken 24 October 2026.

## The TV show that put a fake wristband into AI answers

Every year on 15 March, World Consumer Rights Day, China's state broadcaster CCTV airs the 3.15 Gala, a show that exposes bad treatment of consumers. In 2026, one segment was about GEO.

As [China Daily reported on 15 March 2026](https://www.chinadaily.com.cn/a/202603/15/WS69b6be6aa310d6866eb3de9e.html), an industry insider bought a GEO software system online and made up a smart wristband called Apollo-9. The software wrote more than a dozen promotional articles with invented claims and fake user reviews, then posted them to several self-media accounts with one click. An AI model then repeated the false claims and recommended the device. After 11 more fake articles, including expert reviews and rankings, two AI models listed the wristband among their top picks. The product never existed.

The next day, [Global Times reported](https://www.globaltimes.cn/page/202603/1357055.shtml) that the show described GEO vendors selling paid placement in mainstream AI answers, with offers such as "brainwashing AI" and a "one-week ranking boost". [Yicai reported](https://www.yicaiglobal.com/news/chinas-annual-cctv-consumer-rights-show-uncovers-ai-ad-tricks-that-deceive-customers) that in its own survey at the end of 2025, one GEO provider quoted CNY 3,600 per quarter to optimize a single keyword and promised a place in AI recommendation lists within seven working days.

![China Daily's 15 March 2026 article on AI manipulation of search results at CCTV's consumer rights gala, which describes the fake Apollo-9 wristband test.](https://mandato.agency/assets/img/blog/black-hat-geo/china-daily-gala.jpg)

China Daily's report on the fake wristband test shown at the 3.15 Gala. Screenshot: [China Daily](https://www.chinadaily.com.cn/a/202603/15/WS69b6be6aa310d6866eb3de9e.html), taken 24 October 2026.

## How a scandal became a rulebook in six months

The dates below come from the news reports above and the standard's own page:

- **16 March 2026:** several GEO companies publicly promised to follow the rules and turn away non-compliant work.
- **31 March 2026:** the standard was approved as a project.
- **28 July to 30 August 2026:** the draft was open for public comment.
- **9 and 10 September 2026:** the five parts were issued and came into force.
- **14 September 2026:** they were published on China's national group standards platform.

The rulebook is a set of five voluntary "group standards" from the China Association of Commercial Advertising. It is not a law. [Part 1 of the standard](https://www.ttbz.org.cn/standardDetail/0047401fbeff49148ed25291fc75c245.html) sets out terms and definitions. It applies to brands, GEO vendors, AI platforms and the firms that measure results. It does not cover traditional SEO, model training or ad bidding. The other four parts cover how to assess GEO vendors, pricing and measurement, trusted sources, and compliance in day-to-day service. The drafters include content platforms such as Zhihu, AI and security company 360, and many GEO vendors.

![China's national group standards platform showing Part 1 of the GEO standard from the China Association of Commercial Advertising, marked current, with its September 2026 issue and start dates.](https://mandato.agency/assets/img/blog/black-hat-geo/standard-part-1.jpg)

Part 1 of the GEO standard on China's national group standards platform. Screenshot: [National Group Standards Platform](https://www.ttbz.org.cn/standardDetail/0047401fbeff49148ed25291fc75c245.html), taken 24 October 2026.

## What the rulebook bans

[Part 5 of the standard](https://www.ttbz.org.cn/standardDetail/a7b9efdac5d8402dbfee09bb0fec0704.html) sets rules for the service process, security, risk handling, data protection, content review, record keeping and accountability. The full text needs a sign-in, so we could not read it ourselves. Search Engine Journal, which did, reports that the standard:

- requires contracts to forbid black hat GEO, naming fake content, corpus poisoning, prompt injection, fake traffic and reviews, forged sources and attacks on competitors;
- bans promises such as total control of what a model says or "100% screen dominance";
- requires each contract to include a list of things the vendor cannot promise;
- fails any vendor caught using black hat GEO in a vendor assessment, with a "one-vote veto";
- asks vendors to start fixing errors they placed within 72 hours.

It also says honest advertising, PR, media placements and influencer work are not black hat GEO, as long as they are truthful and lawful. That line matters: the rules do not ban promotion, only lying to the machines that now answer buyers' questions.

![China's national group standards platform showing Part 5 of the GEO standard, on service process compliance and security, with its September 2026 dates.](https://mandato.agency/assets/img/blog/black-hat-geo/standard-part-5.jpg)

Part 5, the part that covers compliance and banned tactics. Screenshot: [National Group Standards Platform](https://www.ttbz.org.cn/standardDetail/a7b9efdac5d8402dbfee09bb0fec0704.html), taken 24 October 2026.

## Why you should not use black hat GEO, in China or anywhere

The rulebook only binds Chinese vendors who follow it. That is no reason to try these tricks elsewhere: they already break rules in other countries, and they can backfire.

- **It breaks Google's rules.** [Google's spam policies](https://developers.google.com/search/docs/essentials/spam-policies), last updated on 28 August 2026, cover scaled content abuse (mass pages made to manipulate rankings), cloaking, hidden text, link spam and more. Sites that break them may rank lower or drop out of results. The same rules guide any good [SEO work](https://mandato.agency/seo-services/).
- **Fake reviews are illegal in the US.** On 14 August 2024, the [US Federal Trade Commission announced a final rule](https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials) that bans fake reviews and testimonials, including ones written by AI. The FTC can seek civil penalties from anyone who breaks it knowingly.
- **It can be exposed in public.** The wristband test ran on national TV. A brand caught faking its way into AI answers loses the trust that GEO was meant to build.
- **It does not last.** AI answers change daily, and platforms keep improving their spam defences. A vendor who promises a spot in AI answers within a week is selling something nobody controls.

![Google's spam policies page at the scaled content abuse section, which describes generating many pages mainly to manipulate search rankings rather than to help people.](https://mandato.agency/assets/img/blog/black-hat-geo/google-spam-policies.jpg)

Google's spam policies already cover mass made content. Screenshot: [Google Search Central, Spam policies](https://developers.google.com/search/docs/essentials/spam-policies), taken 24 October 2026.

## What to do instead: GEO you can defend

Google's own [guide to AI features](https://developers.google.com/search/docs/appearance/ai-features) says the best practices for SEO remain relevant for AI Overviews and AI Mode, and that no special optimization is needed to appear in them. The rulebook's best ideas fit that honest approach:

- **Publish true, useful content.** Pages that answer real questions with facts you can back up are the [content marketing](https://mandato.agency/services/content-marketing/) AI systems quote.
- **Earn real mentions.** Coverage in trusted publications and reviews from real customers count as honest PR under the rulebook, and are the aim of [digital PR for AI search](https://mandato.agency/generative-engine-optimization/digital-pr-ai/).
- **Measure with enough samples.** AI answers vary, so one check proves nothing. Search Engine Journal reports that Part 3 asks for at least 30 runs per question, counting only questions without the brand's name. Our guide to [AI visibility tools](https://mandato.agency/blog/ai-visibility-tools/) covers the free reports from Bing and Google.
- **Put limits in writing.** Ask any GEO vendor for a list of what they cannot promise, and a written ban on fake content in the contract.
- **Pay for the work, not only for placement.** Search Engine Journal reports that Part 3 sets the fixed part of a mixed deal at no less than 30% of its value, so vendors are not pushed to cheat for a bonus.
- **Plan for wrong answers.** Decide who checks what AI says about you and how fast you fix the source when it is wrong. Regular [AI brand visibility checks](https://mandato.agency/generative-engine-optimization/ai-citations/) catch errors early.

![Google's guide to AI features at the SEO best practices section, which says the usual SEO best practices still apply to AI Overviews and AI Mode.](https://mandato.agency/assets/img/blog/black-hat-geo/google-ai-features-seo.jpg)

Google says normal SEO best practices apply to its AI features. Screenshot: [Google Search Central, AI features and your website](https://developers.google.com/search/docs/appearance/ai-features), taken 24 October 2026.

## Where the rulebook falls short

Read the standard with care:

- It is voluntary, and it was written mostly by the vendors it governs.
- The full text needs a sign-in, which makes it hard for buyers outside China to check.
- Search Engine Journal notes that the sample size figures in Part 3 come from a report by the same committee, so treat them as a rough guide.
- It was written for Chinese AI platforms and Chinese regulators. Borrow the ideas, not the exact numbers.

## How to check your own GEO work

Before you publish or pay for anything aimed at AI answers, ask three questions. Is every claim true and backed by a source? Would you be comfortable seeing the tactic explained on national TV? Would it still work if the AI platforms' spam filters got better tomorrow? If the answer to any of them is no, do not use it. If you would like help building AI visibility the honest way, our [generative engine optimization services](https://mandato.agency/generative-engine-optimization/) can help.

[Watch the video summary](https://mandato.agency/assets/video/blog/black-hat-geo.mp4)

Video summary: the six figures from this guide in 36 seconds, with no sound. Everything it shows is also written on this page.

## Frequently asked questions

### What is black hat GEO?

Black hat GEO is any trick that fakes content, reviews, sources or traffic so AI assistants recommend a brand. Examples include mass fake articles, corpus poisoning, prompt injection and invented experts.

### What is the difference between black hat GEO and white hat GEO?

White hat GEO earns AI mentions with true, useful content and real coverage. Black hat GEO tries to fool AI models with fake or hidden material. Only the first is safe to use.

### What happened at CCTV's 3.15 Gala in 2026?

On 15 March 2026, the show reported an insider who used GEO software to post fake articles about a wristband that did not exist. After more fake articles, two AI models listed it among their top picks.

### Who wrote China's GEO rulebook?

The China Association of Commercial Advertising issued it as five group standards in September 2026. Content platforms, an AI and security company and many GEO vendors helped draft it.

### Is China's GEO standard a law?

No. It is a voluntary industry standard. It gains force when companies write it into contracts or when regulators use it as a reference.

### What does the GEO rulebook ban?

According to Search Engine Journal's reading of the full text, it bans fake content, corpus poisoning, prompt injection, fake traffic and reviews, forged sources, attacks on competitors and promises of total control over AI answers.

### Is black hat GEO illegal outside China?

Parts of it already break existing rules. In the US, the FTC's 2024 rule bans fake reviews, including AI-written ones, and Google's spam policies cover mass-made content and link spam. For legal advice in your country, ask a lawyer.

### Can fake content really get a fake product recommended by AI?

Yes. In the March 2026 test shown on CCTV, two AI models recommended a wristband that existed only in fake articles. That is why the rulebook focuses on trusted sources.

### How many times should I test a prompt to measure AI visibility?

Search Engine Journal reports that the Chinese standard asks for at least 30 runs per question. AI answers change between runs, so a single check tells you very little.

### Should I hire a GEO vendor who guarantees AI rankings?

No. Nobody controls what an AI model says. The Chinese rulebook bans such promises and asks vendors to list in the contract what they cannot promise.

## Sources

- [Search Engine Journal: China Wrote The First GEO Rulebook After A TV Scandal (21 October 2026)](https://www.searchenginejournal.com/china-wrote-the-first-geo-rulebook-after-a-tv-scandal-western-seos-should-take-notes/593199/)
- [China Daily: AI manipulation of search results highlighted at consumer gala (15 March 2026)](https://www.chinadaily.com.cn/a/202603/15/WS69b6be6aa310d6866eb3de9e.html)
- [Global Times: AI industry players vow compliance after China's annual 315 Gala uncovers AI data poisoning (16 March 2026)](https://www.globaltimes.cn/page/202603/1357055.shtml)
- [Yicai: China's Annual CCTV Consumer Rights Show Uncovers AI Ad Tricks That Deceive Customers (16 March 2026)](https://www.yicaiglobal.com/news/chinas-annual-cctv-consumer-rights-show-uncovers-ai-ad-tricks-that-deceive-customers)
- [National Group Standards Platform: T/CAACCHINA 001-2026, GEO Part 1, general rules, terms and definitions (published 14 September 2026)](https://www.ttbz.org.cn/standardDetail/0047401fbeff49148ed25291fc75c245.html)
- [National Group Standards Platform: T/CAACCHINA 005-2026, GEO Part 5, service process compliance and security (published 14 September 2026)](https://www.ttbz.org.cn/standardDetail/a7b9efdac5d8402dbfee09bb0fec0704.html)
- [Google Search Central: Spam policies for Google web search (last updated 28 August 2026)](https://developers.google.com/search/docs/essentials/spam-policies)
- [Google Search Central: AI features and your website (last updated 10 December 2025)](https://developers.google.com/search/docs/appearance/ai-features)
- [US Federal Trade Commission: Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials (14 August 2024)](https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials)

[View this as a web story](https://mandato.agency/web-stories/black-hat-geo/)

Last updated: 24 October 2026

---

Mandato Agency · https://mandato.agency/blog/black-hat-geo/ · hello@mandato.agency