Turing Test for Emotional Intelligence (2016)
A traditional Turing Test does a bad job at assessing the intelligent capacities of a computer. The assessment of intelligence depends on whether a human judge thinks, after chatting with both the computer (chat bot) and a human, that the computer is human and the human is a computer. Furthermore, there are simple tricks that can be implemented in the chat bot, which can make it pass the test without any actual implemented A.I. We redesigned the Turing Test. In our version, several chat conversations happen simultaneously, and the human judge is not instructed to find the computer, but in stead is asked to rate the participants on Emotional Intelligence. At the very end the question is asked whether the human judge thinks any of the conversation partners might have been a computer.