From Robert Checchio, Dunellen, New Jersey, US
It isn’t surprising that an AI might recommend nuclear strikes when faced with the prospect of losing a war regardless of the ethical and moral ramifications. Without a soul, and without the ability to internalise the horror and revulsion resulting from the nuclear attacks on Hiroshima and Nagasaki, of course an AI could consider using a nuclear weapon, even if only as a last resort. As long as AIs are guided only by knowledge and experience, not wisdom, ethics and morality, AI-generated terror will be a danger (7 March, p 11)
