The Myth Of Artificial Intelligence by Erik Larson

Why Computers Can't Think the Way We Do

Argues that confident predictions of inevitable human-level machine intelligence are myths: today’s systems excel at pattern recognition and statistical prediction but cannot perform abductive reasoning, generate genuine hypotheses, or understand context as humans do. It critiques hype around deep learning, distinguishes between narrow AI successes and general intelligence, and contends that reaching the latter will require unforeseen scientific breakthroughs rather than scaling current techniques, urging a more realistic, curiosity-driven research agenda.

Purchase from Bookshop.org