I trust artificial intelligence more than the companies that make it. But that’s not what this film is about.
This article is part of the Motherboard Guide to Cinema , a semi-regular column exploring foreign and obscure speculative films.
It’s a truism that technology is morally agnostic. Whether it’s nuclear fission or rocket propulsion, bioengineering or cryptography, it’s the intentions of the people using a technology, not some inherent characteristic with the technology itself, that ultimately determines whether it is good or evil. In this respect, however, artificial intelligence is an unprecedented technology insofar as the predicted end goal of its leading developers is a superhuman intelligence that will likely be beyond human control.
Do You Trust This Computer—a new documentary from Chris Paine, the guy behind Who Killed the Electric Car?—ostensibly set out to map the present state of artificial intelligence with an eye to the ethical consequences of a technology that may one day be beyond human control.
Paine’s film features a who’s who of artificial intelligence: Elon Musk, futurist Ray Kurzweil, IBM Watson’s lead developer David Ferrucci, Open AI’s director Shivon Zilis, Google Brain co-founder Andrew NG, and well over a dozen other academics, authors and CEOs that are all pursuing artificial intelligence in various domains. At just under an hour and a half in length, the film is surprisingly wide in scope, covering everything from robotic surgery and self driving cars to autonomous weapons and humanoid robots. The overarching theme of the documentary is captured adequately in its title, but unfortunately Paine is unable to satisfactorily answer his own question.
The breadth of the film sacrifices any chance of engaging with the question of artificial intelligence and its implications at anything deeper than the absolute surface level. The film mostly comes off as a montage of soundbites from leading AI researchers, who undoubtedly have profound insights to offer about AI, if only Paine had given them the space to do so in his film. The viewer is taken to a handful of research labs in the film, but is only afforded a few minutes at each—enough to do a meet and greet with the artificially intelligent humanoid robot, but not seriously explore its implications with the researchers developing it.
It’s clear Do You Trust This Computer was meant to be a warning about the dangers of AI. The film is dedicated to Stephen Hawking, who was quite vocal about the threat posed by AI. Elon Musk, who is also vocal about the dangers of unregulated AI and co-founded an ethical AI non-profit (which he left shortly thereafter), subsidized a weekend of free streaming. It is surprising, then, that Paine hardly affords his subject any degree of critical distance.
Of course, the pitfalls of AI that we are all familiar with were addressed. Paine considers how Cambridge Analytica used AI to manipulate the 2016 elections; he acknowledges Tay.ai, Microsoft’s chatbot that the internet turned into a Nazi within 24 hours of its release; he depicts the rise of autonomous weapons. At the same time, however, Paine doesn’t push his private-sector subjects to account for how they are ensuring this artificially intelligent hellscape doesn’t materialize.
To take a few examples: Paine gives a lot of screentime to Rana El Kaliouby, a cofounder and CEO of Afectiva, a company that is developing AI to recognize and respond to human emotion. In the film, Kaliouby is shown visiting an elementary school and teaching first graders how the AI works by having them smile or frown into a camera. What isn’t mentioned at all is that Afectiva is developing this technology in order to more effectively target consumers and sell them more Coca-Cola.
Paine also interviews Sean Gourley, the CEO of Primer, who discusses how AI will displace millions of workers. Yet Paine never probes Gourley about his role in shaping the future of AI. Before founding Primer, Gourley worked on algorithms predicting insurgent activity in Iraq and is now applying these wartime insights to business. Among Primer’s first customers were Walmart and In-Q-Tel, the business branch of the CIA.
While the risks of AI are acknowledged in the film, there is hardly any accountability or criticism of how it is being developed right now by these people, aside from some mild hand wringing about Google and Facebook. Instead, most of the subjects discuss the topic as an inevitability. They discuss what will happen, but what I want to know is why must it happen this way. Aren’t the people telling me how AI can quickly spin beyond our control the very people who are in the best position to do something about it?
Perhaps the reason is that, despite the lip service to ethical AI development and all the signatures on all the open letters, many of these people are first and foremost concerned with making as much money as possible.
It is telling when Stanford professor Jerry Kaplan describes machines as “natural sociopaths” and cites the algorithmic stock market flash crash in 2010 as a prime example of this sociopathy. If the people developing the most cutting-edge AI are doing so to extract the maximum amount of profit from consumers (read: the rest of us), of course things probably aren’t going to turn out that great for the world. Machine learning techniques are often compared to the way a child learns: namely by collecting experiences from the world around it and being guided by a mentor. Yet if these mentors are teaching AI to be the most ruthless capitalist it can be, of course it’s going to be sociopathic. But let’s not pretend that’s “natural.”
The one thing artificial intelligence can’t develop on its own yet is will, or the ability to set goals for itself. This must be determined by its creators, most of whom have a goal of maximizing profit. But that’s not what this movie is about.
Before the credits roll, the screen goes black and displays this message: “The pursuit of artificial intelligence is a multibillion dollar industry with almost no regulations.” One has to wonder why Paine didn’t ask the leaders of this industry that he assembled for the film how effective AI regulation would be accomplished and what their own businesses are doing to prevent the shitty future they all describe.
The real movie about the future of AI, the one Paine didn’t make, starts where this one ended.