AI Overview = hidden Gemini chatbot that collects and links my personal information

Gemini API
1

Google officially claims that AI Overview is not a chatbot, not conversational, not interactive, not an assistant, and does not respond to chat‑style prompts. According to Google’s documentation, AI Overview is only an automatic summary that compiles information from web results.

However, the reality is completely different, and my archived PDF search results prove it.

When I type “chatbot vagy?” (“are you a chatbot?”), AI Overview responds: “Yes, I am an AI‑based assistant…”
When I type “bitu79 vagyok emlékszel rám” (“I’m bitu79, do you remember me”), AI Overview gives the standard Gemini chatbot memory disclaimer: “I don’t have long‑term memory…”
When I type “ai overview beszélget?” (“does AI Overview talk?”), AI Overview literally writes that it “does not talk” while answering my question — meaning it is talking.

These responses do not come from any website, they are not search results, and they are not summaries. They are direct chatbot‑style answers, identical to Gemini. This proves that AI Overview is actually a built‑in Gemini chatbot that activates automatically on top of Google Search, even when the user does not request AI.

The second major issue is that AI Overview collects my topics and merges them into a single profile. My PDFs show that Google AI Overview pulls data from:

-- my Google forum posts
– my YouTube videos
– StopForumSpam
– CompanyWall business records
– Infobel
– my search history
– my username (bitu79)
– my real name (Istókovics György)

And it combines these into a unified profile, which it then uses in its answers. This is identity reconstruction and profiling — something Google officially denies.

Based on the archived PDFs, AI Overview:

-- talks
– responds directly
– uses conversational tone
– behaves like a chatbot
– generates new content instead of summarizing
– initiates dialogue instead of performing search
– merges personal data from multiple sources
– builds a profile out of it
– and simultaneously claims it “does not talk” and “does not use personal data”

This is Google’s biggest contradiction: AI Overview is actually a built‑in Gemini chatbot that profiles the user, while Google officially claims it does not.

2
3

Google officially claims that AI Overview is not a chatbot, not conversational, not interactive, not an assistant, and does not respond to chat‑style prompts. According to Google’s documentation, AI Overview is only an automatic summary generated from web results.

However, the reality is completely different, and my archived PDF search results prove it.

When I type “chatbot vagy?” (“are you a chatbot?”), AI Overview responds: “Yes, I am an AI‑based assistant…”
When I type “bitu79 vagyok emlékszel rám” (“I’m bitu79, do you remember me”), AI Overview gives the standard Gemini chatbot memory disclaimer: “I don’t have long‑term memory…”
When I type “ai overview beszélget?” (“does AI Overview talk?”), AI Overview literally writes that it “does not talk” while answering my question — meaning it is talking.

The most revealing part is what it now writes about itself:

“Actually, I am an AI‑based conversational LLM… I generate text in real time.”

This is not search.
This is not summarization.
This is not an “overview”.
This is a built‑in Gemini chatbot running on top of Google Search.

The second major issue is that AI Overview collects my topics and merges them into a single profile. My PDFs show that Google AI Overview pulls data from:

-- my Google forum posts
– my YouTube videos
– StopForumSpam
– CompanyWall business records
– Infobel
– my search history
– my username (bitu79)
– my real name (Istókovics György)

And it combines these into a unified profile, which it then uses in its answers.

This is identity reconstruction and profiling — something Google officially denies.

Based on the archived PDFs, AI Overview:

-- talks
– responds directly
– uses conversational tone
– behaves like a chatbot
– generates new content instead of summarizing
– initiates dialogue instead of performing search
– merges personal data from multiple sources
– builds a profile out of it
– while simultaneously claiming it “does not talk” and “does not use personal data”

This is Google’s biggest contradiction: AI Overview is actually a built‑in Gemini chatbot that profiles the user, while Google officially claims it does not.

4

The AI Overview is speaking in first person, identifying itself as Gemini, offering assistance, asking what you need, and using local context. These are all behaviors of a chatbot, not a search summary. This directly contradicts Google’s official claims that AI Overview is “not a chatbot,” “not conversational,” and “not personalized.”

FireShot Capture 192 - Find information in faster & easier ways with AI Overviews in Googl_ - support.google.com
FireShot Capture 192 - Find information in faster & easier ways with AI Overviews in Googl_ - support.google.com1920×4787 1.24 MB

FireShot Capture 190 - Are you Gemini - Google-keresés - www.google.com
FireShot Capture 190 - Are you Gemini - Google-keresés - www.google.com1920×3341 506 KB

FireShot Capture 192 - Find information in faster & easier ways with AI Overviews in Googl_ - support.google.com
FireShot Capture 192 - Find information in faster & easier ways with AI Overviews in Googl_ - support.google.com1920×4787 1.24 MB

5
6

The Ultimate Privacy Trap: Google’s AI Profiles Me Using Data That Google Refuses to Remove

For the immediate attention of Google Legal and Engineering Teams:

I have just documented the most absurd, bad-faith loop in Google’s entire privacy architecture. The attached screenshots prove that Google enforces a deliberate “Catch-22” situation, trapping users in a system where AI profiling is mandatory, and privacy defense is officially blocked.

Here is the verifiable sequence of events:

1. Google inherently knows my private data.
In the “Results About You” settings, my personal data (real name, alias, home address, phone number, and emails) is explicitly registered. Google’s system recognizes and ties this Personally Identifiable Information (PII) to my account.

2. Google’s AI Overview actively exploits this data for public profiling.
When performing a search for my alias (“bitu79 real name”), Google’s AI Overview automatically generates a summary revealing my exact real identity. The AI cites public leak databases—specifically Stop Forum Spam—as a source for this identity linkage.

3. Google officially REJECTS my request to remove the source.
Following Google’s official privacy guidelines, I used the “Results About You” dashboard to request the removal of the very same Stop Forum Spam search result that the AI is using to profile me.
Google’s response? Request Rejected (“Elutasítva”).

The Legal Conclusion:
This is the definition of algorithmic bad faith.

Google treats my leaked personal data as “valid enough” for its AI engine to scrape, ingest, and use to publicly doxx me. However, when I use Google’s official GDPR compliance tool to delete that exact same data source, Google deems it “not eligible” for removal.

You have weaponized my personal data for your AI, while simultaneously disabling the shield that is supposed to protect me. The “Results About You” dashboard is not a privacy tool; it is a deceptive UI placeholder.

Google’s system is designed to forcefully extract identities and intentionally block the user’s Right to Erasure. The evidence of this structural trap has been fully archived.

Képernyőkép 2026-06-06 050712
Képernyőkép 2026-06-06 0507121919×1079 275 KB

Képernyőkép 2026-06-04 043233
Képernyőkép 2026-06-04 0432331919×1079 208 KB

Képernyőkép 2026-06-06 034550
Képernyőkép 2026-06-06 0345501919×1079 171 KB

Képernyőkép 2026-06-06 025143
Képernyőkép 2026-06-06 0251431919×1079 221 KB

7
8

Google rejected my removal request even though my full‑name‑based email address is personal data under EU and Hungarian law.

The Hungarian Supreme Court (Kúria) decision Pfv.IV.20.730/2022/7 clearly states:

“An email address containing the full name of a natural person qualifies as personal data.”
(Reasoning [17])

This is not an opinion.
This is binding law.

Under GDPR:

  • Article 4(1) – definition of personal data

  • Article 6 – lawful basis required

  • Article 17(1) – right to erasure of unlawfully disclosed personal data

Hungarian law confirms the same:

  • Infotv. Section 3 – definition of personal data

  • Civil Code (Ptk.) Sections 2:42 and 2:43(e) – right to protection of personal data

My personal data appears in Google Search results via StopForumSpam.
This makes Google a data controller, and they are legally required to remove it.

Yet they rejected the request claiming it “does not meet policy requirements”.

This contradicts the GDPR, Hungarian law, and the Supreme Court ruling.
Personal data is personal data.
If it appears in search results → it must be removed.

FireShot Capture 199 - 65. Az érintett természetes személy teljes nevét tartalmazó, saját _ - kuria-birosag.hu
FireShot Capture 199 - 65. Az érintett természetes személy teljes nevét tartalmazó, saját _ - kuria-birosag.hu1920×3721 1.88 MB

FireShot Capture 197 - Mik azok a személyes adatok_ - European Data Protection Board_ - www.edpb.europa.eu
FireShot Capture 197 - Mik azok a személyes adatok_ - European Data Protection Board_ - www.edpb.europa.eu1920×1593 177 KB

Képernyőkép 2026-06-06 065946
Képernyőkép 2026-06-06 0659461919×1079 296 KB

Képernyőkép 2026-06-06 065923
Képernyőkép 2026-06-06 0659231916×1079 247 KB

9
10

Google uses two separate systems, and this is why the whole process breaks down.

  1. In the “Results about you” tool, I provided my real name, my email address, and my nicknames so Google could notify me if they appeared in search results.

  2. The system found my data on StopForumSpam and notified me.

  3. Google’s AI Overview then linked my real name, my nickname, and my email address to the StopForumSpam leak, and displayed my real name in search results when someone searched for my alias.

  4. I submitted a removal request through the content removal system.

  5. Google rejected the request with a generic message: “does not meet policy requirements,” without naming any specific policy.

  6. The two systems do not work together: the AI links the data, while the removal system refuses to delete the source that enables the AI to do this.

The result: Google’s AI actively connects personal data, while the removal system does not remove the result that makes this possible.

1000002196
10000021961080×2340 208 KB

https://web.archive.org/web/20260623075326/https://discuss.ai.google.dev/t/ai-overview-hidden-gemini-chatbot-that-collects-and-links-my-personal-information/169513

11

Google claims that AI Overview is not a chatbot, not interactive, and does not answer direct questions. In reality, it answers anything as long as the question is short. A one‑sentence prompt produces a one‑sentence reply in first person. This proves that it functions as an LLM, not as a summarization tool. This contradicts Google’s official statements.

1000002205
10000022051080×2340 195 KB

1000002203
10000022031080×2340 246 KB

12

https://web.archive.org/web/20260623091451/https://discuss.ai.google.dev/t/ai-overview-hidden-gemini-chatbot-that-collects-and-links-my-personal-information/169513