tech π¬ AI shouldn't just give answers, it needs to adapt
Research found that over-reliance on AI can lead to worse decisions and loss of expertise in complex human tasks. Computer scientists at Harvard SEAS propose an adaptive AI support system, instead of a one-size-fits-all approach. They developed a reinforcement learning model that dynamically decides how much to help a human user. Online experiments with over 1,000 participants showed this learning model outperformed other AI supports. This research suggests AI design must account for human psychology to truly improve performance. π§