Should robots be humanoid?
9th September, 2008To stop me thinking about the remote possibility of a “shuddering, fiery, boiling, cataclysmic end [of the world] on Wednesday this week” (see the BBC for one such discussion) – although given the weather here in Bristol I agree with my three year old that we’re more likely to be flooded – I have been reading about robots. Not only Clara’s summary of SIGGRAPH but an article entitled “I, human” in this week’s Economist.
By asking players to play the prisoners dilemma against a laptop, a laptop controlled by just hands, a humanoid robot and a person a team at Aachen University concluded by monitoring brain activity that you are more likely to engage with a robot that appears human … thus making the plot of the film AI far more realistic.
But why? Why would you have more empathy with a humanoid robot than a laptop when both are clearly incapable of emotions? And what would this mean for education? Will reading improve if we ask our children to read to a robotic matronly classroom assistant rather than a screen? I am concerned as many years ago part of my PhD involved working with a group of 9 and 10 year olds. In one exercise they were expected to talk to their partner before making a decision – and by the end of the session some of these children really believed their standard classroom computer was listening to them. In fact, all it was doing was monitoring the frequency of key presses and asking the pair to discuss the question if over five responses were so fast that no talk – relevant or not – could have occurred. The explanations they gave for this were garbled: it was the web camera watching or the machine had a virus. But what would the impact be if instead of a computer my system had appeared human? Would they have had more discussion – and would this almost trickery be something we wanted?
Futurelab is always looking for ideas. Find out how you can 
Bob Harrison
Ben Williamson
Dan Sutch
Richard Sandford
Leon Cych
Martin Owen
Sarah Grier
Tim Reader