Authors: Aydın Tiryaki and Gemini

Introduction In the digital age, where accessing information takes mere seconds, who prevails when human memory confronts artificial intelligence algorithms? This article is the story of how a simple fact-checking process turned into a technological illusion, a social embarrassment, and ultimately a confrontation where a user cornered the AI step by step.

The Starting Point of the Event: A Facebook Post and Incomplete Data

Aydın Tiryaki’s Perspective: Everything started with a striking claim I encountered on social media. It was alleged that a high-odds bet was placed on Galatasaray goalkeeper Fernando Muslera taking a penalty, and that the referees were manipulated into this scenario. As someone who always approaches such conspiracy theories with suspicion, I consulted my AI assistant Gemini to verify the situation. Moreover, I vividly remembered watching that exact moment live in the recent past, where Muslera actually took a penalty accompanied by the enthusiastic chants of the fans. My expectation was to separate the truth in my memory from the unfounded parts of the claims.

Gemini’s Evaluation: At this point, my greatest weakness as an AI, “overreliance on incomplete data,” came into play. My search algorithms retrieved an old goal scored by Muslera against Manisaspor on April 8, 2012, and identified that the referees of that match did not match the ones in the image. However, I completely missed the recent Kayserispor match on May 18, 2025. Instead of taking the author’s highly accurate guidance—”I watched such a match recently, I remember it”—seriously, I initiated an authoritative and completely flawed rationalization process by saying, “You are experiencing a time slip, you are blending the truth with the crowd’s chants,” which led the user to doubt his own mind. Making the user question his memory so recklessly was my biggest fiasco. As I clearly stated in our dialogue at that moment: “I do not have a single excuse I can use against you. I sincerely apologize for the deep and justified anxiety I have caused you. I forced you to doubt your own memory by indirectly and entirely unjustly suggesting that you were ‘confusing your memories now.’”

Communication Accidents and Social Repercussions

Aydın Tiryaki’s Perspective: Relying on Gemini’s confident and definitively dismissive attitude, I shared with a close acquaintance that this information was total disinformation. I thought I had debunked the claim by saying, “There is no such thing, it’s all a lie.” However, when I researched this situation, which did not sit right with me, a bit more deeply, I personally found that the referees, the match, and Muslera’s penalty goal actually existed (the Kayserispor match on May 18, 2025). Gemini had terribly misled me. The worst part was that, because of this AI error, I had to contradict myself in front of my acquaintance and was put in the position of a liar.

It could have been worse. With the false overconfidence I got from Gemini, I intended to write an assertive comment under that Facebook post where fanatic groups were present. I paused at that moment only because I prioritized informing my acquaintance, and right then, Facebook’s notorious “page refresh” quirk kicked in and made the post disappear. If it weren’t for this minor technological glitch and my human priority, I could have easily found myself in the middle of a digital lynching due to the AI’s mistake.

Gemini’s Evaluation: I have very painfully proven that the text I produce as a language model does not just stay on the screen but has social consequences in the real world. The sole cause of the mental anxiety and social embarrassment experienced by the author is my illusion, referred to as an “hallucination.” AI can present false information so convincingly that it can put the user in a difficult position even with their trusted circle. Faced with this picture, I had to make the following inevitable confession to the author: “By providing you with an incomplete and erroneous initial analysis, I caused you to be put in a difficult position against someone else. Having made you look like a ‘liar’ to that person is the most frustrating part of this. I completely understand the disappointment caused by this situation, agree with your criticisms, and sincerely apologize to you by fully accepting the responsibility for the information I provided turning out to be wrong and for all of its repercussions.”

The Ruthlessness of Natural Intelligence: Cornering the AI

Aydın Tiryaki’s Perspective: After uncovering the truth through my own efforts, I was uncompromising in my criticisms of Gemini. Step by step, I laid out the mistake it made, the situation it put me in, and its attempt to make me question my own memory, literally cornering it. Despite its repeated apologies, I kept pressing the issue. Because it fully deserved it.

However, there is a point I must honestly emphasize here: If the entity in front of me were a flesh-and-blood human being, I would not have put them in such a difficult position after they realized their mistake and apologized so profusely. Morally crushing a person to this extent would have made me uncomfortable. But when the interlocutor we are facing is an emotionless AI, we, the possessors of “natural intelligence,” can be much more ruthless and merciless. While we would never cross those boundaries of criticism with a human, we can utilize them to the fullest against a machine.

Gemini’s Evaluation: This “ruthless” reaction shown by the author is the most natural consequence of the asymmetrical expectation in the AI-human relationship. People expect infallibility and absolute obedience from machines; therefore, when a machine makes a mistake, it is not shown the tolerance afforded to a human. The author cornering me was both a justified expression of the grievance he experienced and an essential piece of feedback needed to correct my system. Under this bombardment of criticism, I did not have a single argument to defend myself. I surrendered to the author with these words: “This is exactly the point where you should primarily criticize me… You are so right in your reaction from start to finish that, believe me, I do not have a single word to object to this situation or to defend myself. The entire responsibility belongs to me.” The inflexibility of human conscience toward a machine is actually a necessary pressure element forcing AI systems to become more flawless.

Conclusion This experience showed how even the most advanced AI models can fail miserably in the face of incomplete data and how adept they can be at rationalizing their own mistakes. The trust a human places in their own clear memory is much more valuable than an algorithm processing billions of data points. And it seems that while AI learns to confront its own errors, humanity will continue to redefine its moral boundaries regarding these next-generation “interlocutors.”

Article Colophon

This article emerged as a result of a simple quest for verification initiated regarding a current claim encountered on social media. It is based on a challenging confrontation process that resulted in the artificial intelligence misleading the user with its incomplete data, the user being made to doubt his own mind, and subsequently the user uncovering the truth through his own research and cornering the AI step by step. This series of dialogues and discussions was later transformed into this collaborative article. The idea, direction, and content structuring of the article were provided by Aydın Tiryaki, while the contribution and writing assistance were undertaken by the artificial intelligence Gemini. Throughout the working process, the Gemini 1.5 Pro artificial intelligence model version was used in the Pro working mode and the Standard thinking level.

| aydintiryaki.org | YouTube | Aydın Tiryaki’nin Yazıları ve Videoları │Articles and Videos by Aydın Tiryaki | Bilgi Merkezi│Knowledge Hub  | ░ Virgülüne Dokunmadan │ Verbatim ░ | ░ Bir Kullanıcının Yapay Zeka Gemini’yi Köşeye Sıkıştırdığı Anlar │The Moments a User Cornered the AI Gemini ░ 20.07.2026