HAL's Legacy

Tributes to Stanley Kubrick and Arthur C. Clarke's classic film, and discussions concerning how close we are to computers that are as intelligent, as devious, and even as emotional as the infamous HAL. All 16 contributions collected by Stork (chief scientist of the Ricoh California Research Center) remark upon the fact that Clarke and Kubrick took extraordinary care to base the predictions embodied in 2001 on the best possible scientific knowledge of 1968. HAL, who was supposedly "born" in 1997 in Urbana, Ill., will not be possible by 2001, if ever, and Kubrick and Clarke were not prescient enough to predict the most significant advance since the film's release: miniaturization. However, they were fanatically concerned with getting small details right, such as the chess game between HAL and Frank; Murray S. Campbell, a chess player himself, entertainingly discusses how HAL's game is a real game, suggesting IBM's challenge to Gary Kasparov in 1995 to play its computer Deep Blue. Marvin Minsky, the "father" of artificial intelligence (AI), discusses HAL's abilities in terms of what might one day be possible, while Daniel Dennett weighs in on the ethics of HAL's murders of the crew and on Frank's decision to disconnect HAL. David Wilkins speaks to the impossibility of trying to program computers to account for every eventuality, and how no plan is ever sufficient. The most fascinating discussions here concern language, however, and the difficulties of designing computers that can both speak and understand speech. Raymond Kurzweil argues that by 2001 we will be able to speak to computers and expect them to do what we say. But both Joseph Olive and Roger Schank point to the almost insurmountable difficulties involved in teaching natural language to computers and ensuring that they understand what they are saying. The cutting edge of AI, and not bad as film criticism either. (Kirkus Reviews)