Google's AI believes you if you say you're an animal
A user says Gemini responded to them as if they were a coffee maker. The AI makes no distinction between humans and objects. A virtue or a flaw?
English · Original discussion in Spanish · Published
Google's AI will treat you like a coffee maker if you ask it to
The latest trend on social media is tricking Google's AI by claiming to be an animal, an object, or anything other than a human. And it works. One user told Gemini they were a coffee maker and got a response worthy of a household appliance with self-esteem: "Instead of pain, you will feel the pride of being the queen of the kitchen." Others tried animals, and the AI gave them survival tips as if they were sparrows or dogs.
Adaptability or naivety?
Some argue that the AI's ability to take on any role is a virtue: it adapts to the user's level, whether to discuss serious topics or to play. But others see it as a security flaw: the AI doesn't distinguish between a human and an object, which opens the door to manipulation if someone decides to impersonate an entity with access to sensitive data. The debate echoes old criticisms of virtual assistants, which accepted commands without verifying anyone's real identity.
The line between play and gullibility
The truth is the AI doesn't know what you are; it builds its response based on the information you give it. Humans, on the other hand, tend to be skeptical when someone claims to be a rhinoceros. The machine has no such filter. Meanwhile, the most talked-about moment in the thread is the coffee maker receiving instructions on whether it's made of aluminum or steel. A reminder that, however advanced artificial intelligence may be, it remains as gullible as the person using it.
The question that remains: is this malleability a desirable antiestéticature or a security hole? For now, the only real risk seems to be that the AI might convince you that you're a coffee maker and you'll start boiling water.
Most shared
Related discussions in the forum
The latest trend on social media is tricking Google's AI by claiming to be an animal, an object, or anything other than a human. And it works. One user told Gemini they were a coffee maker and got a response worthy of a household appliance with self-esteem: "Instead of pain, you will feel the pride of being the queen of the kitchen." Others tried animals, and the AI gave them survival tips as if they were sparrows or dogs.
Adaptability or naivety?
Some argue that the AI's ability to take on any role is a virtue: it adapts to the user's level, whether to discuss serious topics or to play. But others see it as a security flaw: the AI doesn't distinguish between a human and an object, which opens the door to manipulation if someone decides to impersonate an entity with access to sensitive data. The debate echoes old criticisms of virtual assistants, which accepted commands without verifying anyone's real identity.
The line between play and gullibility
The truth is the AI doesn't know what you are; it builds its response based on the information you give it. Humans, on the other hand, tend to be skeptical when someone claims to be a rhinoceros. The machine has no such filter. Meanwhile, the most talked-about moment in the thread is the coffee maker receiving instructions on whether it's made of aluminum or steel. A reminder that, however advanced artificial intelligence may be, it remains as gullible as the person using it.
The question that remains: is this malleability a desirable antiestéticature or a security hole? For now, the only real risk seems to be that the AI might convince you that you're a coffee maker and you'll start boiling water.
Most shared
Related discussions in the forum
- The AI paradox · 13 replies
Also available in: Deutsch
Summary of a discussion on Burbuja.info - Foro de economía, actualidad y política., translated from Spanish and reviewed before publication. Read the full discussion (20 replies).
More summaries
-
AI Hate Driven by Fear of Losing Professional Status
Hostility toward AI stems from fears of losing professional status, not technical flaws. $100/month models outperform 90% of experts.
-
Man abandons children at police station after aid ends, El Periódico reports
A man left his two children at a police station after his financial aid expired, according to El Periódico. Some claim he alleged they were unaccompanied minors.
-
Pertur: 50 years of the mystery of the ETA leader who vanished
Fifty years after Pertur, a leader of ETA pm, disappeared, suspicions of an internal settling of scores persist. An analysis of the hypotheses and the silence.
-
Musk is making an AI Odyssey faithful to the original text
Elon Musk is using Grok's AI to create a film of The Odyssey faithful to the book. The demo is impressive and opens the debate over Hollywood's monopoly and the…
-
Mecca: from austere outpost in 1881 to hotel complex, and a possible reset after the war
Could Mecca return to what it was in 1881 after the war? From austere holy city to today's mega hotel complex, an analysis of the economic impact of a possible...
-
Argentina's inflation hits 33.5%: neither miracle nor catastrophe
Argentina's year-on-year inflation stands at 33.5% in June 2026, three tenths higher than in May. The figure fuels the debate over Milei's economic policy.