The finishing touch

When a person tells their significant other, “I know more about you than you know about yourself”, they are, at best, only broadly correct; the rightful claim, however, seemingly rests with artificial intelligence. In a recent case, AI, in the hands of a group of musicologists and computer scientists, has completed Ludwig van Beethoven’s 10th Symphony – the one that was still in the works when he died in 1827. The symphony is now set to be released in October. Certainly, AI learnt about Beethoven possibly more than the great musician himself could have.

The project is being overseen by Professor Ahmed Elgammal, Director of the Art and Artificial Intelligence Laboratory at Rutgers, the State University of New Jersey; Elgammal had, incidentally, earlier used AI to create “music in the style of Bach”. The project promises to open up endless possibilities of completing the unfinished works of great minds. If a computer, input with Beethoven’s works and a smattering of sketches of the 10th Symphony left behind by the composer, could guess what he might have composed, then would it similarly finish, say, PG Wodehouse’s ‘Sunset at Blandings’?

Elgammal has also developed a program, christened AICAN, which is an ‘autonomous artist’ capable of producing artworks. One of these sold for $16,000 at an auction. For the program, Elgammal relied on a theory proposed by psychologist Colin Martindale that holds that artists typically reject existing and familiar forms, subjects and styles, in order to be refreshingly new. Accordingly, the program will not create anything that resembles an established style.

“As a scientist, I created the algorithm, but I have no control over what the machine will generate,” says Elgammal.