Newscientist reports that roboticists are studying ways for future domestic robots to better interact with humans by sensing their moods. This could be done through biometric sensors, such as for galvanic skin response , or by facial-recognition software equipped with mood analysis algorithms . This would permit your robot to know when its help is or is not wanted.
One problem is that attempting to program a robot for every possible combination of moods and situations would be nearly impossible. Instead, some researchers propose allowing each robot to learn appropriate responses on its own. For example, one robot might learn that gloomy humans appreciate being cheered up, but another robot may learn to leave such humans alone.
In this sense, a robot would become rather like a pet, learning from– and adapting to– its environment Unfortunately, some pets are abused, to often-tragic results. What, I wonder, would be the result of robot abuse?