In at least one of my currently circulating research statements I propose that in the future we might construct cooking robots with a degree of motivational autonomy. Such a robot would control the development of its own tastes, managing them in a way responsive to our reaction – the feedback we give it – but without just cooking to our taste. It would develop tastes of its own with a degree of autonomy!
I find this an fascinating case to consider, but one might sensibly raise the question why anyone would build such a robot? Why endow a cooking machine with such an autonomy of taste rather than just make it cook according to our tastes? We might do it, because the developing taste intrigues us. We do not only consume the food on our plate as a pile of nutrients and spices. We have an interaction with its producer. The autonomy of the robot becomes part of the experience. The higher-developed the autonomy of taste, while remaining response to our feedback, the more we have to discover in this other agent.
I take this to be a vaguely Hegelian thought: One of the key experiences we have as human agents is living in the presence of other autonomous agents and enjoying it. Interaction with highly developed artificial intelligence might at some point transform this experience, in ways we have yet to grasp. Understanding the potentials of autonomy in artificial systems and the different types of autonomy, the topics of one of my research streams, will be a step forward.
 I won’t discuss here how far into the future we are looking with this scenario. To answer that question a longer discussion of the nature of taste would be required.
 I am purposefully only writing of a degree of autonomy. There is more to autonomy than what I am sketching here.