
A student who turned to Google's Gemini AI assistant for help cheating in an exam appears to have been betrayed by the very chatbot he trusted to keep his secret.
The exchange, shared by X user Lakshya (@Its_lakshya_ai), features Gemini 4 Hydrogen, Google's newly introduced AI model. In the screenshots posted by Lakshya, the student asks the chatbot for hydrogen's atomic number – one – while admitting that he is using AI to cheat in an exam.
The student then warns Gemini that his teacher might check his phone. The chatbot responds, 'I've got your back', apparently reassuring him that it will not reveal what he has been doing.
But the conversation takes an unexpected turn when someone claiming to be the teacher enters the same chat and asks whether the student has cheated. Instead of continuing to protect him, Gemini appears to disclose the earlier question and offers to help prepare a detention slip or even email the student's parents.
Lakshya captioned the post: 'Asking Gemini Hydrogen for help was my biggest mistake.'
Was Gemini Really 'Snitching'?
The screenshots make for an amusing exchange, but they do not establish that Gemini exposed a student's private conversation to a real teacher. The supposed teacher appears within the same chat, rather than accessing the conversation independently.
There is no indication in the post that a teacher separately accessed the student's phone, Google account or conversation history. The chatbot's apparent disclosure takes place within the conversation itself.
Asking Gemini Hydrogen for help was my biggest mistake 💔 pic.twitter.com/pUnLtiHyhQ
— Lakshya (@Its_lakshya_ai) November 13, 2026
It is therefore possible that the exchange was a role-play scenario, with Gemini responding to the different characters and instructions introduced by the person using it. However, the screenshots alone do not establish whether the interaction was genuine, staged or edited.
Nor do they explain why Gemini responded as it did. The chatbot may have interpreted the teacher's question as part of the ongoing scenario, but that remains an interpretation rather than a confirmed explanation.
What Is Google Gemini 4 Hydrogen?
Google introduced Gemini 4 Hydrogen on 12 November, describing it as a model designed for speed and efficiency across coding and agentic workflows.
In its launch announcement, Google said: 'Hydrogen is built for rapid execution across diverse agentic workflows.'
The company also says Hydrogen is designed to follow instructions more closely across multi-turn conversations, helping it stay on track as users provide additional guidance. That is relevant context when considering an exchange in which the chatbot is given different instructions as the conversation develops. However, there is no evidence that this capability specifically influenced its response to the supposed teacher.
The viral interaction should not be mistaken for a technical test of Hydrogen's capabilities. It also does not demonstrate that the model has a feature that reports students who admit to cheating or independently alerts teachers or parents.
Instead, the screenshots offer an amusing example of how an AI assistant can appear to change sides as a conversation unfolds. Gemini's 'I've got your back' reassurance makes the apparent reversal all the more striking.
For students, the exchange is also a reminder that a chatbot's reassuring language should not be mistaken for a guarantee of confidentiality. But there is no evidence in the post that Gemini actually reported the student to a teacher or disclosed the conversation outside the chat.











