Science Mistakes in “Ex Machina”
This past weekend I watched the movie Ex Machina. I liked it quite a bit, but as a computer scientist I was naturally interested in the technology that appeared in the film. This post outlines my thoughts on it.
Note: This post contains no spoilers. The basic plot can be seen in the trailer).
First off, artificial intelligence (A.I.) in real life is an entire discipline. Robotics — how a machine is constructed, and how it senses and processes information — is its own area of study. Even the recognition and synthesis of natural languages (such as English) is a huge field.
So it’s a stretch to say that a single person could build an advanced android by himself. No matter how smart he is, it’s practically impossible for someone to become an expert in all these fields and have time to actually construct and test it single-handedly (and have plenty of time to hit the gym). It takes a team to build amazing stuff.
The movie also slightly misrepresents A.I. and the Turing test. It talks about A.I. like it is a goal (“you’ve achieved A.I.”), rather than continual progress (like any field). The initial description of the Turing test is mostly correct, but then they go through the test sessions which (as the character points out) is nothing like a Turing test. Performing a “Turing test” when it is obvious from the start that the subject is an android doesn’t make any sense. I can see why the movie uses the term — “Turing test” is fairly well-known and can be distilled to the simple idea of “robot passes as human”.
Most perplexing of all is how a computer genius seems to have never heard of the 3 Laws of Robotics. Isaac Asimov defined them more than 70 years ago, and they have been used in movies for almost as long. So it seems odd that the scientist constructing a near-perfect android would not try to incorporate them (or similar safeguards). Even if there is some flaw or philosophical reason behind it, he shouldn’t be surprised when she becomes dangerous.
Luckily, unlike big-budget Hollywood films, the technology makes enough sense to tell the story. The jargon used by the characters fits the conversations. Some of the stuff they mention is rooted in current science (such as determining if a person is lying by facial and body cues). There is no outright ridiculousness as in Prometheus or Lucy. The “wetware” brain of the android is the closest thing to pure sci-fi — it’s plausible, but requires a jump in current technology for it to be possible.
So in all, the film is much more about philosophy than science; more about humanity than robots. The science is explained just enough to serve the story. Which makes it a pretty good sci-fi movie.