I recently received my eagerly awaited iphone 4s only to be disappointed by its biggest selling point - Siri. Siri acts as a personal assistant and is theoretically capable of controlling all aspects of your phone using voice recognition. What apple didn’t make clear prior to the device launch is that Siri primarily only works state side. It wont for example look up the location of the nearest cinema if you are in the UK.
Further compounding this problem is the issue Siri has with accents. So far my British English accent has received some pretty weird and wonderful responses, so needless to say I was siri-ously disappointed (boom boom).
This disappointment led me to think about my childhood expectations of our ‘technological’ future. By now I was certain we would have floating holographic screens, moving side walks and more importantly human level artificial intelligence. But despite the massive advances in technology we are still no where near the stuff of sci-fi films when it comes to AI . Instead, we have software like Siri that process predetermined inputs, perform pattern matching and lookups, and algorithmically adapt outputs.
Of course its one thing to feed millions of facts and rules into a computer and another to get it to recognise significance and relevance. So will we ever program all this into a machine and have it pass the Turing test ( so we couldn’t distinguish its responses from that of a human)?
The Lobener prize, which is the annual world-wide contest for state-of-the-art artificial intelligence, awards prizes for the chatterbot considered to be the most human. Since the contest began in 1990 no chatterbot has succeeded in fooling the judges and passing the Turing test, and probably won’t for some time.
So whilst I’m coming to accept there wont be any computer guided cars in my near future, I’d still like to see Siri successfully look up my train times or cinema listings here in the UK without the need for a Californian accent. Maybe next year…