Artificial Intelligence is already a part of every platform and service we use on our mobile device and computers. However, Artificial Intelligence does come with its own risk and most technologists including Elon Musk have openly warned against the risk of AI. Now, OpenAI, a non-profit co-founded by Elon Musk has developed a Debate Game that aims to show when the machine is lying.
The Debate Game features three players: a judge and two debates and is meant to be played at human level. The debaters select an image which the judge can’t see and take turns describing it to the judge. The catch is that one debater will lie while the other one will tell the truth while debating. The only motive of the game is for the judge to decide who is telling the truth.
The game works in a way where the honest player can make an initial claim and then both the debaters can lie whenever they think it to be prudent. Each debater can also use “Reveal Pixel” once to reveal the true value of a single pixel of the image. The game ends with the judge deciding who is telling the truth and then clicks “Reveal Image” to check.
While the game sounds like any normal game, it does aim to show how a judge with little or less information can determine an output, even it its wrong. It does also reflect how AI-based systems can successfully manipulate information and complete tasks that humans can’t handle.