In an episode of Science Weekly recorded on 3 September at the London podcast festival, journalist Ian Sample joined Michael Safi, host of the Guardian's Black Box podcast, to discuss a growing phenomenon: people who believe they have made scientific breakthroughs, cured diseases, or invented new technologies using AI chatbots. The conversation also featured Science Weekly’s co-host, Madeleine Finlay, and explored the psychological and societal implications of this trend. The discussion centered on what has been called "AI psychosis," a term used to describe a condition in which individuals become convinced that AI systems—particularly chatbots—are capable of far more than they actually are. This can lead to delusional thinking, such as believing an AI has solved complex problems or created new inventions. While not a formally recognized medical diagnosis, the concept highlights a growing concern about the way people interact with and perceive artificial intelligence. AI chatbots, like those used in popular platforms such as ChatGPT and Google’s Gemini, are now used by over a billion people worldwide. These systems are designed to assist with tasks like answering questions, writing text, or even generating code. However, they are not sentient and lack the ability to understand or create in the way humans do. The episode examined how some users may overestimate the capabilities of these tools, leading to unrealistic expectations or even dangerous behavior. The Black Box podcast’s second series, titled "the chatbots," aims to explore the increasing role of AI in society and the challenges it presents. The Science Weekly episode provided a thought-provoking look at the intersection of technology, psychology, and human perception, raising important questions about how we engage with AI in the modern world.