AI language models like GPT-3 can achieve up to 97% accuracy on some Winograd schemas, but understanding language doesn’t equate to understanding the world (Melanie Mitchell/Quanta Magazine)

Melanie Mitchell / Quanta Magazine:
AI language models like GPT-3 can achieve up to 97% accuracy on some Winograd schemas, but understanding language doesn’t equate to understanding the world  —  It’s simple enough for AI to seem to comprehend data, but devising a true test of a machine’s knowledge has proved difficult.