In a huge step for the field of artificial intelligence (AI), the first computer program was successfully able to trick a human into thinking they were speaking to another human.
This test – known as the Turing test after its creator Alan Turing – requires a human judge to be listening to a conversation between two subjects, one of which will be an artificial intelligence with a with another human, both of which are also separated and cannot see one another.
In this case, the judge was unable to say conclusively that one of them was a form of AI and therefore marked the first instance of a program passing the test.
The Russian-made AI named Eugene Goostman was able to convince one of the three judges tested during the experiment that the AI was in-fact a 13-year-old Ukrainian boy who had a non-native understanding of the English language.
According to io9, the test was part of a competition within the Royal Society of London that pitted five AIs against one another to see if any would be able to finally pass the Turing test.
In order for an AI to pass the test, it must be able to fool the human judges at least 30pc of the time as to whether it is human or AI, and, in this case, was able to convince them 33pc of the time.
Not quite Skynet
However, critics of the results have downplayed the seemingly monumental leap-forward in AI that has been made out by this news.
The writers of io9 believe that Eugene’s creators perhaps knowingly chose the model of a non-native young person so that any imperfections or seemingly non-traditional forms of expression could simply be passed off as expected of a child of that age, something which is “not in the spirit” of the test.
Another problem arises from the fact that Eugene is not a supercomputer or advanced form of AI, rather it is a ‘chatbot’ computer program that works of pre-programmed scripts rather than naturally responding to conversation.
However, Prof Kevin Warwick from the University of Reading had attended the event and spoke of the test as a significant milestone in eventually truly accomplishing real AI: “Of course the test has implications for society today. Having a computer that can trick a human into thinking that someone, or even something, is a person we trust is a wake-up call to cybercrime.
“The Turing test is a vital tool for combatting that threat. It is important to understand more fully how online, real-time communication of this type can influence an individual human in such a way that they are fooled into believing something is true…when in fact it is not.”
AI image via Shutterstock