ÁñÁ«ÊÓƵ

Futurologist Ian Pearson ("Wisdom received, over and out", 18 June) suggests that "more rapid reverse engineering of brain processes" coupled with an alleged exponential progress in technology will deliver conscious machines by 2020.

In 1962, Arthur C. Clarke, in Profiles of the Future, predicted that we would see artificial intelligence by the mid-1990s: 47 years later, we're still nowhere near it. The central problem remains one of giving machines real-world semantics so that, like humans, they can elaborate rich models of reality.

Thus, it is baffling that understanding natural-language semantics is not given the same international priority as the Human Genome Project. It is even more baffling that this prospect is being further retarded by the curtailment of university linguistics programmes.

Greg Michaelson, Edinburgh.