Prof Geoffrey Hinton, the “godfather of artificial intelligence”, states that he struggles to find examples of “more intelligent thing[s] being controlled by … less intelligent thing[s]”; the mother-baby relationship is the only example he can cite (‘Godfather of AI’ shortens odds of the technology wiping out humanity over next 30 years, 27 December). This seems a strange outbreak of aspect blindness, especially given Hinton’s specialism.
Many theorists (Graham Harman, Timothy Morton, Jane Bennett, Bruno Latour and others) offer persuasive arguments showing how (to borrow from Freud) “man is not master in his own house”: human behaviour is continually, at times conspicuously, regulated by non-human drivers, many of them seemingly pretty dumb. Coronaviruses offer a topical example. The present barely regulated rise of AI is unarguably scary, but dealing with it effectively will involve humans getting real about their non-mastery of all they survey and interrogating the ways that stuff (both smart and dumb) controls us, as well as vice-versa. The same goes for climate breakdown and ecological crisis.
Rachel Withers
London
Prof Hinton says there are few examples of a “less” intelligent being controlling a more intelligent one. Leaving aside the difficulty of actually defining what intelligence is, can I suggest that the age-old and still-thriving institutions of slavery and political oppression fit the bill? Any differences in intelligence between the oppressors and those they oppress is irrelevant as to how successful these enterprises are. Of course, if we deliberately hand over the management of our affairs to these benighted AI machines, we deserve everything we get.
George Burt
Glasgow
