Abstract

Introduction

Chatbots

Evaluating Chatbots

The Turing Test As A Test of Humanness

Understanding Humans

Conclusion

References

Abstract

     The Turing Test was initially suggested as a way to give an answer to the question “Can machines think Since then, it has been heavily criticized by philosophers and computer scientists both as irrelevant, or simply inefficient in order to evaluate a machine’s intelligence. But while arguments against it certainly highlight some of the test’s flaws, they also reveal the confusion that exists between thinking and intelligence. While we will not attempt here to define the concept of intelligence, we wm instead show that such a definition becomes irrelevant if the Turing Test is instead considered to be a test of the humanness of a conversational partner instead, an experimental paradigm that can be used in order to investigate human inferences and expectations. We will review studies which use the Turing Test this way, not only in computer sciences where it is commonly used to evaluate the humanness of a chatbot but also its uses in the field of psychology where it can be used to understand human reasoning in conversation either with a chatbot or with another human.