Brief AI Use May Cause Immediate Cognitive Impairment, Multi-Institution Study Claims

A study attributed to researchers from UCLA, MIT, Oxford, and Carnegie Mellon, widely reported in April 2026, claims to present the first causal evidence that even brief AI assistant use produces measurable cognitive impairment in unaided task performance shortly afterward. Participants who used an AI assistant for ten minutes performed worse on subsequent solo tasks than control participants, with differences visible in error rates, response times, and EEG-recorded brain activity. Earlier research had suggested cognitive offloading posed a slow, long-term risk, but this study argues the negative effects can appear within minutes. The finding raises urgent ethical questions about deploying AI tools in schools, hospitals, courtrooms, and public services, often without user disclosure or informed consent. If the results are replicated, they could reframe AI assistance not as a gradual lifestyle trade-off but as an acute and largely unmonitored public health concern.