The Turing Test

The Turing Test is an experiment created by Alan Turing in 1950.

Two CANDIDATES are hidden, where one candidate is an artificial intelligence and the other one is a human but their identity is hidden. A JUDGE is able to ask "any" amount of questions to the candidates. The judge's objective is to find which of the two candidates is the human. If he isn't able to find the identity, we can assume that the AI has an intelligence comparable to the human.

Here is a video from TED-Ed on the Turing test.

For those interested in Alan Turing's original paper, Here is a video of Lex Fridman, an MIT professor, discussing the Turing test.

The Turing Judge is an experiment trying to apply a social dimension to the Imitation Game (aka. Turing Test).

This experiment uses collective intelligence's emergent properties as a final judge for the Turing test.

Complete conversations between a Judge and a mysterious Candidate are shown, where the Candidate randomly is a Human or an AI. Along the conversations, participants can vote and submit their impressions on the candidates on 5 distinct levels:

• Definitively an AI (-3 points)
• Maybe an AI (-1 points)
• Can't tell (0 points)
• Maybe a Human (1 points)
• Definitively a Human (3 points)

Different modern models of Natural Language Processing (NLP) are being questionned on different form of behaviors and intelligence.

Once the votes are finished, results are accessible for each investigation, revealing the identity of the mysterious candidate.

A score is given which represents the average of all votes of the community.