AI Reading Claims Need Stronger Evidence Before They Shape Enterprise Decisions

Claims that an AI model learned to read efficiently and independently adopted human-like behaviours, such as skipping predictable words or revisiting difficult text, should not yet shape enterprise AI