ChatGPT Shows Human-Like Flaws in Math Study

A study by Hebrew University and Cambridge researchers found ChatGPT exhibits unpredictable, learner-like behavior and makes human-like errors when solving math problems.

By Sarah Jenkins