In 1950 Bell Labs scientist, mathematician and WW II cryptographer Alan Turing wrote a famous paper posing the question “Can machines think?” He devised a hypothetical test in which an Interrogator poses written questions to two unseen Subjects, one human and one machine. Based on the Subjects’ responses, the Interrogator then tries to determine which is the human and which is the machine. Turing predicted that, by the year 2000, machine Subjects would be capable of fooling 30% of human Interrogators into thinking that they were the human.


Dr. Alan Turing 1912-1954

The problem for this “Turing Test” was that, for many years, there was no practical way to conduct it. Computers were just not that far advanced. So the Turing Test remained a purely  academic thought experiment until, in 1991, a highly successful entrepreneur marketing pre-fabricated disco dance floors, Hugh Loebner, put up a substantial cash prize sponsoring an annual Turing Test (called the Loebner Prize, naturally). In addition to the cash, the winner is awarded a Gold Medal and the title “The Most Human Computer,” which I can only imagine would move the machine to tears of joy and pride.

The Loebner event is replete with computer contestants (human imitator “chatterbots”), a panel of interrogators (“judges”) and human volunteers as subjects (“confederates”). Although Turing’s year 2000 prediction has not quite been realized, the machines are now getting quite close, fooling 25% of judges in recent testing. In order to motivate the human subjects to act more like humans (thereby making it harder by contrast for the bots to appear human) the organizer also gives a prize each year to “The Most Human Human.” Winner of the award Brian Christian has written an interesting book and highly entertaining presentation  (the talk starts after five minutes of introduction which might well be skipped).

Every year since 1966 another award, the Association for Computer Machinery (ACM) Turing Award, is given to the best overall contribution to the field of computer science. This most prestigious award is essentially the Nobel Prize of computer science. In 1983, Ken Thompson (below left) and Dennis Ritchie (right) of Belle Labs won the Turing Award for the invention of UNIX.

Computer Expert Swindlebeards
Swindlebeards Thompson and Ritchie

If you have any interest in computer security, Thompson’s brilliant Turing Award lecture is a must read –  Reflections on Trusting Trust: “To what extent should one trust a statement that a program is free of Trojan horses? Perhaps it is more important to trust the people who wrote the software.”

Among Thompson’s other pioneering contributions (rambling back to chess) he is noted for work with Joe Condon on the Computer Chess World Champion Belle (1980), which included, by the way, four- and five-piece endgame tablebases solved by retrograde analysis (working backwards from final positions rather than forward from present positions). Thirty-plus years ago this was ground-breaking stuff. Nowadays if you’re having trouble sleeping I’d recommend a visit here or here or just google  “pawnless endgames fifty move rule.”

Here is an early game played by Belle.

[pgn height=550 initialHalfmove=16 autoplayMode=none]

[Event “Computer WCh”]
[Date “1978”]
[Round “?”]
[White “ACM BLITZ 6.5”]
[Black “Belle”]
[Result “0-1”]
[ECO “C48”]
[PlyCount “28”]

1. e4 e5 2. Nf3 Nc6 3. Nc3 Nf6 4. Bb5 Nd4 {The Rubinstein Variation} 5. Bc4 (5.Nxe5 Qe7 6.f4 Nxb5 7.Nxb5 d6 8.Nf3 Qxe4+ 9.Kf2 Ng4+ 10.Kg3 Qg6!) 5…Bc5 6. Nxe5 Qe7 7. Bxf7+ $2 (7. Nxf7 O-O $13 {Yes, White’s play has been weak, but consider that this was 34 years ago and only a few years after invention of the first microprocessor.}) 7… Kf8 $17 8. Ng6+ $4 hxg6 9. Bc4 Nxe4 $19 10. O-O Rxh2 $19 {Crunch! Other moves also win but this is the most human …} 11. Kxh2 Qh4+ 12. Kg1 Ng3 13. Qh5 {White sees the mate now and plays to prolong the game by one move} 13…gxh5 14.fxg3+ Nf3# 0-1 {Simultaneously blocking the check while administering double check and mate!}

[/pgn]