While we are not quite there yet, roboticists from the University of Pisa have brought us even closer with FACE - a robot who is able to display very lifelike human emotions, ranging the whole gamut from happiness to sadness and fear.
Building on 30 years of research, the robot uses 32 motors in 'her' face to simulate different expressions, and in doing so brings us closer to the 'uncanny valley' - the point where a robot looks almost realistic enough to pass for a human, but robot enough to freak us out.
The university team, led by PhD student Nicole Lazzeri, modelled the robot on one of the team's wife - and the team say 'it's very realistic'.
The team uses software called HEFES to Hybrid Engine for Facial Expressions Synthesis, which has been built up over the last 30 years to mimic the emotional responses of people.
It tells the motors how to respond in order to recreate a look, and the software can mix the various emotions on a sliding scale, for instance a smile tinged with sadness, or a laugh mixed with unease.
This last expression might be apt for those people who get trapped in the 'Uncanny Valley'.
The principle of the valley is that people do not have an issue with humanoid robots when they do not look human - for instance, when they have metallic faces or expressionless faces.
And if a robot is undetectable as a robot, then the illusion is not broken, and people can interact with the robot.
But if the robot looks very close to human, yet is given away by tell-tale signs, it makes most people startled and very uneasy - which, if you plot on a graph of human reactions to robots, create the 'uncanny valley'.
According to New Scientist, the research team asked five autistic and 15 non-autistic children to identify a set of expressions performed by FACE, and then the same expressions performed by a psychologist.
Both groups could identify happiness, anger and sadness, but they struggled more with fear, disgust and surprise.