“Risk with a Robot – Are you game?”

A review of “Improving Social Presence in Human-Agent Interaction” – CHI 2014

“Humans have a tendency to consider media devices as social beings”

 The above sentence, the first of the paper, certainly catches your attention. In the world we live in, we interactive with devices on a daily basis. Often, we know these devices do not have real emotions or ideas and yet still attribute social presence to these devices.

The paper I’ll be reviewing is from CHI 2014. It attempts to experimentally demonstrate that by following a set of predefined guidelines (see below), it is possible for a social robot, the agent, to have an improved perceived social presence. The paper refers to the Frank Biocca definition of social presence which can be described as “the sense of being together with another”.

To improve social presence, an artificial board game opponent should:
1.    Have a physical embodiment and be able to engage in face-to-face interaction with participants
2.    Be believable and obtain players’ attention
3.    Have an emotion system to make better judgements and simulate human emotions
4.    Have social memory
5.    Be able to simulate social roles common in board games

 For the experiment, a social robot plays the Risk board game against 3 human players. The hypothesis posed is whether the agent that follows guidelines for socially present board game opponents will be perceived as more socially present than an agent that doesn’t follow the guidelines. Therefore, the independent variable is whether the agent demonstrated the guidelines or not.

Risk was a good choice of game to use as it is social and causes the players to take various social roles. The agent used was an anthropomorphic robot called EMYS which could speak, gaze at opponents and have a memory of past experiences with players.

 Questionnaires were used to collate results from participants. To better measure social presence, the authors broke the concept down to six dimensions. This allowed the authors to ask questions that focused on these dimensions individually. Further general questions were asked to better understand what participants thought of EMYS. The results showed when the agent followed the guidelines, it was perceived as more socially present.

“I felt like he was a real companion that socialised with all of us”

The authors of the paper put a lot of effort into making the agent seem believable. A question I would pose though is, is this what people want? Do we want to live in a world where we interactive with our devices in just the same way we interact with each other? Another paper from CHI 2014 titled ‘Mobile Attachment - Causes and Consequences for Emotional Bonding with Mobile Phones’ shows that we already have emotional connections to our personal devices; yet they are not anthropomorphic.

I’d also like to see a further experiment following from this one where the players cannot see each other and therefore do not know they are playing with a robot. Would they interact with the device as if it were human?

I will come back to this topic in further blog posts as I explore interactive devices and how humans use them. I’ll leave you with this thought. Can we truly stay connected through technology?