Latest in Science

Image credit:

'Ex Machina' shows Turing isn't enough to test AI

Shares
Share
Tweet
Share
Save

With Ex Machina, the directorial debut of 28 Days Later and Sunshine writer Alex Garland, we can finally put the Turing test to rest. You've likely heard of it -- developed by legendary computer scientist Alan Turing (recently featured in The Imitation Game), it's a test meant to prove artificial intelligence in machines. But, given just how easy it is to trick, as well as the existence of more rigorous alternatives for proving consciousness, passing a test developed in the '50s isn't much of a feat to AI researchers today. Ex Machina isn't the first film to expose the limits of the Turing test, but it's by far one of the most successful. And, like the films 2001 and Primer, it's a work of science fiction that might end up giving you a case of philosophical whiplash.

Ex Machina is constructed like a morality play. A nebbish programmer, Caleb (Domhnall Gleeson), wins a Willy Wonka-esque contest to spend a week with Nathan (Oscar Isaac), the reclusive genius CEO of his company -- a search giant called BlueBook. Nathan, who is equal parts startup bro, Steve Jobs and Larry Page and all with a megalomaniacal bent, arrives with an intriguing dilemma: He's developed an artificially intelligent robot named Ava (Alicia Vikander) and he wants Caleb to prove she has a consciousness of her own. Of course, they're both well aware that the Turing test alone won't be enough.

As originally conceived, the Turing test involves a natural language conversation between a machine and human conducted through typed messages from separate rooms. A machine is deemed sentient if it manages to convince the human that it's also a person; that it can "think."

Since it was developed, there have been claims of several AI experiments that have successfully passed the Turing test. When subjected to scrutiny, however, these results were found to be invalid. Just last year, a chatbot pretending to be a 13-year old boy convinced 33 percent of a judges panel that it was human -- effectively passing the Turing test's threshold of 30 percent believability. But it wasn't long before a slew of critics, including the long-running technology-analysis site TechDirt, pointed out that its programmers had gamed the test by pretending the chatbot wasn't a native English speaker -- allowing it to take advantage of a handicap -- and that plenty of other chatbots have achieved better scores.

No matter an AI's final Turing test score, a script built to imitate human conversation or recognize patterns isn't something we'd ever describe as being truly intelligent. And that goes for other major AI milestones: IBM's Deep Blue is better at chess than any human and Watson proved it could outsmart Jeopardy world champions, but they don't have any consciousness of their own. It's worth noting that neither of those supercomputers has gone through the Turing test, though inventor and futurist Ray Kurzweil believes Watson could be retooled to pass it easily.

In Ex Machina, Caleb has nothing like Blade Runner's Voight-Kampff machine, a souped-up lie detector used to detect artificial life, so he has to come up with more creative ways to prove Ava's intelligence. He's also not dealing with a mysterious entity in another room; he's talking to her face to face through a window, and can clearly see that she's a robot. All of that makes it more difficult for Ava to trick him like a typical chatbot, although it opens up new ways for her to influence him. The sexual politics of the situation aren't an accident: Ava's a strangely attractive robot created by a hyper-masculine genius CEO, who finds her Prince Charming in Caleb. And you could argue that it's also a cheap way for Nathan to manipulate Caleb's investigation.

Tim Tuttle, a former MIT AI researcher and the CEO of the predictive-intelligence company Expect Labs, agrees with Garlands' use of sexuality in the film as a ruse. "[Ex Machina] proposed a sort of inverse where it's not enough to have a human be deceived for a machine to be real," he said. "The machine needs to convince the human to do things for it -- to fall in love with it, to serve its own purposes."

There's another element present in the complex conversations Caleb has with Ava and one familiar to consciousness researchers: the Chinese Room Argument. It's a thought experiment philosopher John Searle formulated as a way to disprove the usefulness of the Turing test. The basic gist of which goes like this: You're stuck in a room following along with a computer program that's replying to questions being slipped under the door in Chinese. By following the computer's instructions, you manage to output perfect responses in Chinese characters. But even so, you clearly don't understand the Chinese language or what those responses actually mean. Similarly, Searle argues a computer doesn't actually need to understand language or complex thought to pass the Turing test. And it's this line of thinking that colors Caleb's interactions with Ava throughout the film. He continually goads her to prove she understands what it is she's saying.

Ex Machina also touches upon the more useful offshoots of the Turing test like Lovelace 2.0, which focuses on having machines demonstrate true creative abilities. It's not much of a spoiler to say that, on several occasions, Ava demonstrates that she can be just as creative as any human.

We never learn how, exactly, Nathan invents Ava, but it's clear his search engine BlueBook has played a big role. At one point during the film, he notes that most people thought BlueBook existed just to give them answers when, in fact, it actually served as a means of collecting those queries and creating a framework for artificial intelligence. That may sound like pure sci-fi fluff, but according to Tuttle, "People inside Google really feel that way." It's just not something the general public recognizes.

If search giants like Google believe a fully aware AI is an inevitability, and an offshoot of our own archived curiosity, just how will we end up testing its existence? Ex Machina, for one, doesn't provide that elusive answer, but it certainly raises a lot of the right questions. And right now, we're mainly in the realm of thought experiments when it comes to judging consciousness. But, at the very least, the film makes it clear that we need to think beyond simplistic tests like Turing and, most importantly, we should be prepared to be surprised.

[Photo credits: DNA Films/Film 4]

'Ex Machina': The Turing Test Clip

From around the web