This article was originally posted on RealClearScience.
Warning to reader: There are *SPOILERS* in this movie review.
Ex Machina, a suspenseful new film about artificial intelligence, is disturbing for all the right reasons.
Many of those who work in the field of artificial intelligence believe that, someday, AI will advance to the point that a computer or robot will be indistinguishable from a human. Any AI that accomplishes this feat will be said to have passed the Turing Test, named after Nazi codebreaker and ingenious British computer scientist Alan Turing.
When, exactly, this will occur is a matter of debate. Ray Kurzweil, an infamously optimistic futurist, believes that robots will essentially be calling all the shots in about 30 years, an event referred to as thesingularity. How is this optimistic? Because Mr. Kurzweil also believes that, by the year 2045, we will be able to upload our minds into some sort of device that will allow our consciousness to live forever. So, perhaps it might be worth investing in some extra thumb drives.
Don’t bother, Tom Hartsfield, our curmudgeonly physics blogger, would say. Not only do Mr. Kurzweil and other AI groupies improperly use the mathematical term singularity, but the entire idea of super-advanced AI borders on being physically impossible. As the soon-to-be Dr. Hartsfield explains, singularities (think: infinities) do not occur in nature. There is no such thing as something (in this case, artificial intelligence) growing at an exponential rate for all eternity. Eventually, it hits a limiting factor.
The reason, for instance, that bacteria and bunny rabbits have not overgrown the planet is because they run out of food, produce too much waste, and/or get eaten by something bigger. Similarly, the reason, according to Mr. Hartsfield, that we will not produce an infinitely intelligent AI is because we will run out of something, be it silicon or energy or rare earth metals. Besides, Mr. Hartsfield glumly notes that we can’t even figure out how to model a worm’s brain, with its pathetic 302 neurons and 6,000 synapses. The notion that we are even remotely close to modeling a human brain, which sports roughly 86 billion neurons and 100 trillion synapses, is sheer absurdity.
Bummer. Party pooper.
Thankfully, Ex Machina doesn’t wade into this debate. It just assumes that an impressive AI will occur in the not-too-distant future, whether it be 25 or 500 years from now. Most technologists would probably agree that, singularity or not, exceptionally lifelike robots will come around sooner or later.
Ex Machina begins with Caleb, a clever programmer working for a search engine company, being flown by helicopter into a remote, undisclosed location hours away from civilization in order to meet his boss, Nathan. A tough-looking, reclusive alcoholic, Nathan invites Caleb to his secretive compound for the express purpose of analyzing a robot named Ava. Specifically, he wants Caleb to determine if Ava passes the Turing Test.
Caleb and Nathan begin by probing the basic questions: What counts as passing the Turing Test? What is consciousness? What is a real emotion? Can a real emotion be programmed? Caleb tries to discern answers by talking to Ava in multiple interview sessions.
That’s when the film takes a more ominous turn.
Nathan ends up being considerably darker and weirder than Caleb suspected. Ava rather quickly becomes flirty with Caleb. While tantalized, Caleb is deeply unsettled. How can a robot have sexuality? Why would Nathan program the robot in such a way? Elements of seduction, deception, and manipulation appear — all traits that are unmistakably human.
At the beginning of his visit, Nathan asks Caleb to answer the question, What do you think of Ava? However, by the end of his visit, it becomes increasingly clear that Caleb should have spent more time pondering, What does Ava think of me?