Cluely, an AI app that boldly claims to help you "cheat at everything, " deserves some credit for its honesty. If you view artificial intelligence as a platform that enables gaming the system—by generating smooth prose via a large language model (LLM), creating stunning AI art, or producing convincing videos—then Cluely and its founders are simply transparent about their intentions. More likely, however, they are bluffing their way toward launching a product that undermines integrity. Cluely is a clever browser-installed app that secretly watches and listens to your screen activity. Its main use is during Zoom video interviews for high-paying developer jobs, where candidates are often asked to solve coding challenges like Leetcode problems in real-time. With Cluely active, users receive quick responses and solutions, effectively cheating during the interview. Lee, the co-founder, claimed to have used Cluely successfully in interviews with major tech firms. This reportedly led to his expulsion from Columbia University, as shared in letters he posted on X (formerly Twitter). Remarkably, Lee showed little remorse, instead appearing satisfied with the outcome. Despite controversy, Cluely has raised $3. 5 million in funding and is poised to help users "cheat at everything. " This bold claim may unsettle many. Lee has defended the app on X, and Cluely’s tongue-in-cheek launch video depicts him using the tool—via smartglasses—to deceive a date about his age and interests, only for her to realize the deceit and walk away. The video illustrates how Cluely could even aid in cheating at dating. Though Cluely has not disclosed which LLM powers its answers, Lee’s now-removed manifesto argues for redefining cheating, likening Cluely’s function to that of Google or calculators—tools that deliver answers at our fingertips and have arguably "helped us cheat" for decades. When asked on X whether his stance risks teaching the wrong lesson about AI, potentially increasing skepticism and distrust during job interviews, Lee replied: “What exactly does ‘cheat on everything’ mean?

How do you ‘cheat’ on a conversation?Cluely (and AI) enables such insane leverage that it feels unfair and like cheating. The world will be uncomfortable with this reality, but it will soon become normal. When AI maximalism is normalized, humanity’s potential will increase exponentially. Initial hesitation followed by massive adoption has been the pattern for all technologies. ” While hesitancy followed by widespread adoption is historically accurate, Lee’s claim that you can’t "cheat" in a conversation overlooks fundamental deception—something illustrated in works like Cyrano de Bergerac. This deception is embedded in Cluely’s design. Further probing on X highlighted that while Google and calculators are transparent tools of knowledge and computation—often restricted to prevent cheating—a browser tool that secretly monitors conversations and generates answers without the other party’s knowledge is fundamentally different. Lee has yet to clarify which LLM Cluely uses, perhaps because the answer underlines the ethical concerns. In essence, Cluely is not simply a tool; it is a covert aid that pretends the user is answering on their own. This is the equivalent of secretly using Google or a calculator during an interview, only far more concealed and deceptive. I personally tested Cluely on a MacBook Air, where it required user permission to bypass typical safeguards that prevent third-party apps from accessing video and audio streams. It does not alert other participants that AI is actively watching and listening. The tool is not fully real-time—users must click an AI button for responses—but its potential for misuse is clear. Lee’s attempts to normalize and redefine cheating neither lessen the deception nor make it more acceptable. After being expelled for misuse, he now encourages others to risk their reputations by using Cluely—a step I am unwilling to take, and one many will likely reject. In summary, Cluely offers a powerful yet ethically questionable way to cheat in situations requiring genuine knowledge or interaction, casting a shadow over trust and integrity in AI-assisted activities.