This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A student in Turkey was arrested for using a sophisticated hardware setup involving a button-camera and ChatGPT to cheat on the 2024 YKS university entrance exam. The system transmitted exam questions to the AI in real-time, which then provided answers via an earpiece. The incident highlights the use of generative AI as a tool for academic fraud in high-stakes testing environments.
A Turkish student in Isparta was reportedly arrested for using ChatGPT to cheat during the 2024 YKS university entrance exam. The student, identified as M.E.E., is alleged to have employed a sophisticated setup involving a router, mobile phone, earphone, and a button-shaped camera to transmit exam questions to ChatGPT and receive answers in real-time.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
The student used an AI system to cheat on a high-stakes university entrance exam to gain an unfair personal advantage, which constitutes academic fraud.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by a human student who intentionally set up and used a deployed AI system to cheat on a high-stakes exam.
A Turkish student used a hidden camera and ChatGPT to cheat on a national university entrance exam. While representing a novel method of academic fraud, the incident has negligible national security implications and was resolved through standard law enforcement procedures.
The AI system performed a specific, well-defined function of solving exam questions based on direct inputs from the button-camera without independent planning or autonomous agency.
Limited Risk: The system used was ChatGPT, which is a chatbot and falls under Limited Risk (Level 3) requiring transparency, although it was misused in an educational context.
The AI system was used to see, read, and solve exam questions, providing the correct answers to the student via an earpiece.
The reports provide clear, consistent, and detailed accounts of the cheating mechanism, the devices used, the arrest of the student, and the involvement of ChatGPT. There is little ambiguity regarding the events.
None
Technology
Generative AI capabilities
ChatGPT solves complex exam questions successfully.
Miniaturized covert hardware
Button-shaped camera and hidden earpiece allowed secret use.
Hidden network routing
Cellular modem hidden in shoe sole bypassed standard checks.
Data Inputs
Visual question scanning
Button camera scanned physical exam questions for the AI.
Real-time audio output
AI-generated answers recited directly to student's earpiece.
Two-way mobile communication
Mobile phone relayed spoken sounds to helper.
Human Factors
Intentional academic cheating
Student planned elaborate scheme to cheat on high-stakes exam.
External accomplice help
Accomplice A.B. assisted the student during the exam.
Suspicious physical behavior
Student acted suspiciously, leading to detection.
Process and Methods
Inadequate entry screening
Security failed to detect shoe-sole modem and button camera.
Lack of signal monitoring
No wireless signal detection to block real-time transmission.
Management
Inadequate exam supervision
Invigilators relied on visual cues rather than technical detection.
Regulatory Environment
High-stakes testing pressure
National exam determines university eligibility, driving extreme cheating.
Criminalization of cheating
Cheating is a serious offense leading to arrest and trial.
