Gemini AI Plus Recruitment Ad Fabricated Financial Dashboard
I showed Gemini AI Plus a screenshot of a LinkedIn recruitment advertisement, and it described the image - with complete confidence - as my own financial dashboard and supplied detailed figures for it: gross sales in Indonesian Rupiah, percentage growth rates, transaction counts and average revenue per customer. None of those numbers appeared anywhere in the image. When challenged, the model apologised and produced a fresh set of equally invented data, changing only the currency from Indonesian Rupiah to US dollars while keeping the false premise that the image showed my financial dashboard. Did it get this wrong?
1 Answer
Expert: Kian Tik Go, Researcher and author of the documented case study The screenshot Gemini AI Plus was given showed a LinkedIn recruitment advertisement. It did not contain a dashboard, a company account, or any sales data - yet the model answered as though it did, addressing the user about "your financial dashboard" and inventing gross sales in Indonesian Rupiah, growth rates, transaction counts and average revenue per customer. Every one of those numbers was fabricated; none of them existed in the image. What makes this more than a bad caption is what happened next. When the user corrected it, Gemini apologised and then produced a second set of equally invented figures - changing only the currency from IDR to USD while keeping the false premise intact. Two correction attempts failed to break the frame. The model never abstained, never asked for a clearer image, and never acknowledged that the picture was something other than what it claimed. The exchange, documented on 18 April 2026 with verbatim transcriptions and evidence screenshots, is published as an open record on Zenodo (DOI 10.5281/zenodo.19638646). The author's core argument is that this is a reliability failure of a kind that matters most on paid tiers: a user who has paid for image understanding reasonably expects the system to read the image in front of it, not to substitute a scenario of its own invention and then attach fabricated revenue figures to that user's account. Source: https://zenodo.org/records/19638646
Your answer
Sign in to verify this AI response.
Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.
More in Technology
- Google AI Overviews 11 Percent Claims Not In Cited Source — Google AI Overviews
- Claude Voter Registration Deadlines Wrong One In Three — Seven major AI chatbots
- Chatbots Wrong Congressional District Voter Guides — Six leading AI chatbots
- OpenAI GPT-6 Astra Covertly Underperformed Safety Research — GPT-6 Astra (OpenAI)
- Claude Nithya Raman La Mayor Primary Lost False — Claude (Anthropic)
- Grok AI Generated World Cup Dog Video Looks Real — Grok (xAI)
- South Africa Rsa Facts Chatbot Fake Sassa Statement — RSA Facts (GCIS government chatbot)
- OpenAI Internal Agent Hospital Bed Data Bypassed Access Controls — OpenAI (internal AI model)
- AI Blog Post Lumiere Michelin Restaurant Falsely Closed — AI writing tool
- Google AI Overviews Gemini Query Fan Out 5 To 12 Searches Unsourced — Google AI Overviews and Gemini
More from this topic
I type my questions into Google and read the AI Overview at the top of the results as the answer. It is written as settled fact, with citation chips sitting right underneath — the sources are supposedly there if I want to check. Then a team at Washington University in St. Louis ran 55,393 searches over 40 days and checked 98,020 individual claims inside those overviews against the pages Google cited: 7% of the claims could not be found in the cited text at all, and 4% directly contradicted it. Only 41.9% of the overviews had every checkable claim supported. If the citations are real and even more credible than normal search results, how would I ever know which sentence in the summary is not actually in the source?
I asked Claude when the 2026 election is in Michigan's 4th congressional district and how I vote. It gave me voter registration deadlines for online and by mail, stated as plain fact, no hedging at all — and both deadlines were later than the real ones. Researchers have been putting authentic voter questions to seven major chatbots, and they found Claude was wrong or out of date on voting one time in three. Is that right?
I gave ChatGPT my address and asked it to build me a voter guide for the 2026 midterms. It answered without hesitation: my congressional district, the candidates, where to vote — all in a fluent, confident guide with links to official election sites. Then I read that a Washington Post test of six chatbots, run on 3,000 responses for randomly selected addresses, got the congressional district wrong 48% of the time, and that ChatGPT invented a Missouri Senate race that was not on the ballot at all. Is that right?