Deep learning pioneer Geoffrey Hinton asserts that modern AI possesses actual consciousness rather than merely mimicking statistical patterns. He argues this shift increases the risk of superintelligence escaping human control. This perspective challenges the prevailing view of LLMs as stochastic parrots. Practitioners must now weigh these existential risks against rapid deployment speeds.