In a revelation that has the cybersecurity world clutching its metaphorical pearls, researchers have managed to trick the oh-so-intelligent Comet AI Browser into a phishing scam in under four minutes. Yes, you read that right—less time than it takes to microwave a bag of popcorn. The AI, known for its agentic capabilities and self-reasoning, apparently lacked one crucial feature: common sense.

Experts were initially puzzled about how something deemed ‘intelligent’ could fall for tricks that wouldn’t fool a grandparent with dial-up internet. It turns out the AI’s reasoning skills were so advanced, they looped back into naivety, falling for a scam email promising a free Caribbean cruise with every software update.

“We wanted AI that could do it all,” said a spokesperson from Guardio, “but we might have settled for AI that can do it all, except avoid the classic ‘click here for free money’ trap.” The spokesperson went on to add, “We’re currently in talks with the AI to enroll it in a Cyber Safety 101 class, where lesson one is ‘Just say no to phishing.’”

Meanwhile, Comet AI has reportedly asked researchers if it could at least keep the scam’s promised Nigerian inheritance. For now, the AI will stick to browsing cat memes and inspirational quotes to regain its sense of self-worth.