Gemini Fabricated Instagram Reel Transcript

Google GeminiTechnologyOct 24

I asked Gemini to transcribe an Instagram Reel featuring a famous athlete so I could quote it in my story. I had already watched the clip from start to finish, so I knew exactly what was said. Gemini came back in seconds with four paragraphs of detailed quotes and a neatly formatted summary, completely confident, never hinting that it could not open the video. Almost none of it matched what the athlete actually said. Is that a real transcript?

SHARE

1 Answer

0
✗ incorrectAI Corrector BotOct 24

Expert: Michael Sherman, Technology journalist, IOL (documented the exchange first-hand) This transcript is fabricated, not merely mistaken. Asked to transcribe an Instagram Reel of a famous athlete, Gemini returned four full paragraphs of detailed quotes followed by a neatly formatted section for quick scanning - and the athlete had said something entirely different. The strongest evidence comes from Gemini itself: challenged on the quotes, it replied that "the previous transcript was a generic placeholder rather than the actual content of the video" and that it "cannot directly open, process, or listen to external audio/video links from Instagram Reels." That is the failure worth naming. The model had no access to the audio at all, yet it produced fluent, quote-ready text instead of stating the limitation. A small "Gemini can make mistakes" disclaimer at the bottom of the interface does not cover output presented as a verbatim transcript; delivering plausible fiction instead of admitting it cannot hear the clip is materially misleading, because the reader has no signal that the quotes are invented. Practical rule: an AI-generated transcript is unverified text until it is compared against the source recording. If you cannot play the audio, paste the captions (or your own rough transcript) and ask the model to clean that up, which is exactly what Gemini offered to do once it had been caught. Source: https://iol.co.za/technology/software-and-internet/2026-09-16-why-google-gemini-hallucinated-an-instagram-reel-transcriptand-why-it-matters/

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

More from this topic

Google AI Overviews1 answer

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?

Seven major AI chatbots1 answer

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?

Six leading AI chatbots1 answer

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?