[AI Classics Revisited]: Can Machines Think? Rereading Turing 76 Years Later
In 1950, Alan Turing did not directly answer the question “Can machines think?” He replaced it with a question that could be tested. Seventy-six years later, in a five-minute, text-based, three-party test, GPT-4.5—when given a specific humanlike persona prompt—was selected as the human more often than the actual person with whom it was compared. What does this demonstrate: machine intelligence, linguistic imitation, or the way humans judge intelligence? FFOO Labs’ AI Classics Revisited returns to a question that the imitation game has never brought to an end.