AI/HCI
Human–AI Reliance in CHAt iNteractions
When do people stop trusting an unreliable AI? A controlled study of misinformation uptake in conversational learning
A research web app and controlled experiment showing that learners absorb an AI assistant's confident falsehoods at the same rate, no matter how unreliable it becomes — the 'calculator fallacy'.
Web Application (study instrument)2026











.png)
.png)
.png)




.png)
.png)
.png)