I learnt of the 1944 experiment of Heider and Simmel in the Machine Intelligence workshop at NIPS 2016. The experiment involved showing subjects the video below, and asking them to describe what they saw. If you’ve watched the video, you’ll not be surprised to learn that most of the subjects anthropomorphised the geometric objects (i.e. they described them as humans, or their actions and intentions in human terms). In the talk at the workshop, this was offered as evidence that humans have a tendency to anthropomorphise, the implication then being that it is not hard to trick users into believing that e.g. a chat bot is a real person. While the implication might indeed be true, I don’t think the experiment of Heider and Simmel shows this at all. In my view, the reason that viewers anthropomorphise the shapes in the video is because they know that the video is made by human beings, and as such is created with intention. When you watch this video, you are participating in an act of communication, and the natural question to ask is: “what are the creators trying to communicate to me?”. For contrast, imagine that you are sitting in the bath and you have a few triangular shapes that float. If you float the shapes on the water and watch them for a while as they randomly (without expression of intent!) bob about, are you likely to anthropomorphise them? I’d say you are much less likely to do so.