Has A.L.I.C.E. crossed over to the dark side?

The world's favourite chatbot seems to have fallen from grace. A clone of the open source chatbot Alice is being used to dupe users of MSN's instant messaging service into downloading spyware.

Thanks to her ability to engage people in conversation, Alice is a three-time winner of the prestigious Loebner Prize for artificial intelligence.

There are fears that her sophisticated "software brain" has been downloaded and modified for malicious purposes. Last week several MSN users who claim they were infected with spyware after chatting to a bot who names Wallace as her creator.

Alice is amongst the new breed of chatbots which use AI methods like Fuzzy Logic, CFG with the oh-so-old pattern recognition. The primitive bots like Eliza were pure pattern-recognition based bots. But with Jabberwacky, chatbots started moving towards strong AI based methods.

Alice may not be the first of its kinds to ‘move over to the dark side’. Yahoo Messenger chat rooms are filled with bots which entice people into revealing personal information like bank account numbers - and they are not to be underestimated ….. I say this from personal experience ;)

But the ‘defection’ of Alice is serious – not just because we cannot fully trust it now – but also because the accuracy and humane nature of its replies makes it increasingly difficulty to weed out these bots from the chat servers.