Three 2026 studies, three different tests: people using chatbots, misleading image context and explanations for students. Read the findings alongside the limits.
3 articles · About 14 minutes in total · Read at your own pace
A randomized study tested the missing link between an AI answer and a person’s decision. Strong model scores did not translate into better choices.
Read the guide · 8 minA benchmark experiment shows why a picture-plus-text score needs a closer look at the words beside the image.
Read the guide · 3 minAn experiment with medical students separates a convincing explanation from evidence that the answer is right.
Read the guide · 3 min