By Toby Murray
Publication Date: 2026-03-12 19:14:00
When OpenAI launched ChatGPT-5 in August last year, many academics scoffed at the tech company’s claim that its new artificial intelligence (AI) model had “PhD-level intelligence.”
For how could systems so prone to hallucinations, faulty thinking and sycophancy compete with the brightest young minds in the world?
But now academics routinely use tools like ChatGPT to help them with their research in the same way they might once have done with graduate students. Perhaps the most famous example is the world’s best-known mathematician Terence Tao, who reports using generative AI as a mathematical collaborator.
I myself recently transformed from a skeptic to a believer when, over the course of a few months, I conducted a research project in which I used a series of generative AI tools to perform tasks that I would normally perform in collaboration with my graduate students.
But this experience has also highlighted a hidden danger of AI – one that shows why it…
