Share

On April 18, 2025, Delhi Police announced the dismantling of a highly sophisticated, technology-driven cheating racket that targeted government recruitment examinations. The alleged mastermind, a firefighter employed with the Chandigarh Fire Department, used a mobile command center—a converted car—to transmit AI-generated answers in real time to candidates inside exam halls via concealed nano earbuds and customized Bluetooth devices.
The case came to light after the Central Sanskrit University filed a complaint on July 26, 2024, concerning unfair practices during a Lower Division Clerk (LDC) recruitment exam held at a Kendriya Vidyalaya in Delhi’s Janakpuri. During frisking, officials recovered a custom-modified mobile phone from a female candidate and a SIM-card-equipped Bluetooth device with hidden nano earbuds from another candidate, Hemant. A total of 213 candidates vied for just six vacancies.
According to police, the syndicate charged aspirants between ₹10 lakh and ₹12 lakh (approximately $12,000–$14,400) per candidate for the illegal service. The package included specially modified gadgets—Bluetooth devices, nano earbuds, and custom phones—that allowed the candidate to receive answers without detection. The answers were generated using AI tools like ChatGPT and relayed from outside the examination hall.
The investigation began with the arrest of the two candidates. Hemant disclosed that his brother Rinku had arranged the equipment through a contact named Virender for a payment of ₹12 lakh. Police arrested Rinku and Virender, who then led them to Sanni, a woman friend of the arrested female candidate. Sanni had provided the modified phone, Bluetooth device, SIM card, and earbuds in exchange for ₹10 lakh.
Sanni, upon questioning, named Deepak as an associate of the network. Deepak, a high school graduate (Class 12), allegedly arranged the electronic gadgets and coordinated the real-time transmission of answers during exams. Through technical surveillance, social media analysis, and intelligence inputs, police traced Deepak to a larger operation directed by Sonu Kumar, alias Sonu Maan, the suspected kingpin.
Sonu Kumar was tracked to Charkhi Dadri in Haryana, where he was caught red-handed while facilitating cheating in another government recruitment exam. Police recovered a tablet linked to a foreign account, customized mobile phones, Bluetooth devices, nano earbuds, SIM cards, and other electronic gadgets, along with incriminating documents.
The modus operandi described by police reveals a level of sophistication unprecedented in India’s exam fraud history. Sonu had converted a car into a fully equipped mobile control room, complete with a laptop, high-speed internet, multiple mobile phones, and a power supply. While candidates were sitting for the exam, Sonu—parked outside the test center—would receive the questions through a live feed from a hidden camera or a screen-sharing app installed on the candidate’s phone. He then used AI tools like ChatGPT to generate answers within seconds, which were relayed via Bluetooth to the nano earbud worn by the candidate.
“This syndicate represents a new frontier in exam cheating, where artificial intelligence and consumer electronics are weaponized to bypass security,” said a senior Delhi Police officer, speaking on condition of anonymity. “The use of a modified car as a command center is particularly alarming—it shows that the perpetrators were willing to invest heavily in infrastructure to ensure success.”
The case has sparked concerns about the growing use of AI in fraudulent activities, not just in India but globally. In the United States, similar schemes have been reported in college admissions exams and professional licensing tests, though rarely with such a coordinated, multi-stage approach. Cybersecurity experts note that the barrier to entry for such scams is lowering as AI tools become more accessible.
“The same technology that powers ChatGPT can be used to generate answers in real time. The only difference is the ethical boundary,” said Dr. Rebecca Lin, a cybersecurity researcher at Stanford University. “As long as there is high-stakes testing, there will be attempts to exploit AI. This case should serve as a wake-up call for exam authorities worldwide to adopt AI-specific detection tools, such as behavioral analysis and device fingerprinting.”
Police have arrested seven individuals, including Sonu, Deepak, Sanni, Virender, Rinku, and the two candidates. Further investigations are underway to identify additional beneficiaries and the full extent of the syndicate’s network. Authorities are also analyzing the financial trail—the payments made to the gang, the foreign-linked account found on Sonu’s tablet, and whether any exam officials were complicit.
The Central Sanskrit University has announced a review of its security protocols for future exams, including the use of metal detectors, signal jammers, and AI-based proctoring software. “We are also examining the possibility of canceling the results of the affected exam and may re-conduct it if necessary,” a university spokesperson said.
This case underscores a broader trend: the convergence of AI and low-cost electronics is enabling cheating at a scale and sophistication previously unseen. While the arrests are a significant win for law enforcement, experts warn that without systematic upgrades to exam security, similar schemes will continue to emerge.









