Imitation could be a way to solve particular problems in human computer interaction. AI uses machine or deep learning which is applied for complex problem solving to learn in the same way humans do through trying and learning with feedback loops. But if we look into research (which is still at early stages) we know that there appears to be the existence of special mirror neurons – brain cells which copy responses of other humans even if the action was not done themselves. It could be the part of the ingredients for empathy and why we seem to feel things done to others so powerfully. When we copy other people’s facial cues or change our language to fit into the conversation, that could be the mirror neurons at work.
So given this, is there a place in AI for imitation of a human rather than learning from them?
Instead of trying to ‘calculate’ an emotion, the machine can take the human’s lead and just ‘mirror’ their emotion. For positive emotions this could be very effective – we know humans respond well to smiles or positive reinforcement but what could be the potential problems?