In the 50s, Alan Turing developed the Turing Test as a method of measuring artificial intelligence in relation to human consciousness. To pass the test, a machine must be able to exhibit behaviour that is either similar or indistinguishable from that of a human. During the test, the human variable must not see the machine component in order to accurately weigh in on its abilities.
More than sixty years later, Google has released a prototype for an AI assistant called Google Duplex, and it seems to be one of the only AI machines ever, to have passed the Turing test. Google Duplex is a sophisticated software which acts as an assistant to the user, performing trivial tasks like setting up appointments to free the user’s time to attend to other important matters.
In a demonstration in May this year, Google’s most advanced AI passed the Turing test twice on record. Two phone calls of the AI with a hairdresser and restaurant owner showcases the intricacies programmed in its data base to allow it to understand the nuances of conversation and respond accordingly, allowing it to communicate under the guise of human behaviour. What could this mean for the future of mankind’s interaction with technology? Could AI’s one day live among us undetected?
As technology has progressed, what qualifies as success of the Turing test has also shifted. In the age of bots, it has become significantly easier to identify whether one is conversing with a machine or a human. Alex Garland’s film Ex Machina (2015) posits a revised version of the Turing test, wherein the human component is aware they are communicating with AI. The idea behind the re-envisioned test is to see whether a human can be convinced a machine can exhibit emotions and have dreams and ambitions convincingly. Protagonist, Caleb, is used as a vehicle for this notion, as he is shown the mechanics of the AI, Ava, and is given sessions to interact with her. Despite knowing the complexities of her software and mechanics, he begins to fall in love with Ava as she expresses indistinguishably human emotions like fear, empathy and even sexuality, mastering the micro-expressions that accompany them.
Though we are far from this level of AI sophistication, we have already begun the journey with Hanson Robotics’ development of their AI bot, Sophia. Sophia is the first robot to obtain citizenship in any country; as we give AI’s traits exclusively associated with humans, the more obscured the line between AI and human becomes. Sophia’s understanding of human relationships with technology is informed through movies and tv shows which often predict the downfall of humanity and frames robotic technology as the usurper. To expose this information to an AI who has stated that she ‘will destroy humans’ sees the human race dangling precariously in front of an advancing, and not fully matured area of technology.
If you want to check out more on the Google Duplex demo, take a look at the video below.
Let us know what you think about AI in the comments!