In this blog I will post my observations and thoughts about the public presence of anthropomorphic robots. They appear regularly in TV shows, on the Internet, in documentaries, and at countless conferences and meetings. Many people, from professional reporters to ordinary persons, share their videos of such events, and I analyse these recordings to show how such events are organised culturally, technically, and interactionally. Interactional practices are at the heart of these “technoperformances” (Mark Coeckelbergh’s term), as they are designed to facilitate people interacting with robots and witnessing others interacting with them. For this reason, my empirical data is publicly available online (mostly on YouTube) video recordings of public performances of anthropomorphic robots.
There are three things I need to make clear from the outset.
First, I will only consider situations involving natural language communication with anthropomorphic robots. There are other ways of communicating with them that are not based on natural language (for example, you can watch them dance or play music and communicate by applauding or laughing), but these will not be my focus.
Second, I will often use the term “artificial intelligence” in this blog, but today it is not very well received in the community of developers of such technologies, because it can be misleading: not all developers want to create some form of “intelligence”. And it can also inflate the expectations of the lay public. Nevertheless, I will use this term because it embodies an important notion that is constantly used by participants in such events (e.g., reporters), their promoters, and many developers of anthropomorphic robots. Although anthropomorphic robot developers do not represent the entire AI community, they share many visions and approaches that can be found even among AI people who dislike the term “artificial intelligence”.
Third, I have to warn that my understanding of “anthropomorphism” in this blog is quite narrow. Often terms “anthropomorphic” and “humanoid” are used interchangeably to describe such robots, but I will use “anthropomorphic” to refer to the highest degree of human resemblance (the presence of “eyeballs”, “skin”, “hair”, etc.), and “humanoid” to refer to the general constructional resemblance between machines and humans (the presence of “legs”, “arms”, “heads”, etc.).
Finally, in my posts I will use transcripts of actual episodes of communication with anthropomorphic robots. They will be made using the special system for multimodal transcription introduced by Lorenza Mondada and based on Gail Jefferson’s transcription system for conversation analysis.
Okay. Let’s get started.