My first attempt at conversation w GPT-2. It evades questions, flunks basic arithmetic, and gets caught up in probabilities of blank spaces.

Conversations with bots like Eliza and GPT-2 can look decent for short periods, but fail if you insist they keep track of anything. pic.twitter.com/ZVNBewtu4J

— Gary Marcus (@GaryMarcus) November 12, 2019

Exactly what I’ve noticed. So-called AI can pass the Turing Test (or variants thereof, yes I know how the original version of the Turning Test was configured) only if the person testing the AI is very stupid.