Groundbreaking research exposes critical flaws in using large language models to detect AI-generated academic work, showing they struggle to identify human writing and can be easily deceived—throwing academic integrity safeguards into question.