My chief purpose in this paper is to direct attention towards a problem which has received little attention from philosophers, but which I think has a greater claim to importance than any other problem that would naturally arise from a consideration of robots and cybernetic machines: both because it seems of greater intrinsic philosophical interest, and because it represents more nearly than any other problem the kind of worry that afflicts ordinary people. This problem may be roughly formulated in the question ‘What can men do that machines can't ?’