Unlike Skynet or HAL 9000, Johnny Depp makes for a particularly suave and seductive artificial intelligence in the new film Transcendence. Directed by Wally Pfister, the film posits a near future where A.I. researchers are on the verge of achieving “the Singularity,” that highly anticipated moment where computers will surpass humanity’s capabilities and create systems of near infinite wisdom and power. But one thing appears to be missing: human consciousness, which, through a series of unfortunate events, Depp’s scientist Dr. Will Caster, eventually provides.
Filled plenty of thoroughly unrealistic technologies, Transcendence nevertheless taps into current scientific around brain-machine-interfaces and nanotechnology. Science and Film spoke with the scientific advisors on the movie, Jose M. Carmena, an associate professor of electrical engineering and neuroscience at UC Berkeley, and Michel Maharbiz, an associate professor of electrical engineering and computer sciences at Berkeley, about monkey brains, cyborg beetles, and how little we know about the human mind.
Sloan Science and Film: How did you get involved in the film as scientific advisors?
Michel Maharbiz: The filmmakers went to a number of places: MIT, Stanford, and Berkeley’s Eric Fraser passed them onto me. I quickly realized that bringing in Jose was the right thing to do. Their team really wanted to understand this stuff, which was remarkable, because most of the time filmmakers just want a rubber stamp. So we read scripts, and edited it line by line, and then went down to L.A. and sat with Wally and his group for ten hours. And we went through it again line by line. Wally would ask us questions, and then his art and special effects people would run out and make something and show us. And we’d say, “Yes, that’s good.”
SSF: What do you think were the most crucial contributions you made to the film?
Jose Carmena: There were two major contributions: one has to do with the neuro-technology used by Johnny Depp to upload his mind, and other one relates more to the nanotechnology in the second part of the movie. The neuro-tech part of the narrative would require opening up Will’s skull, so we had to devise a compromise between something that looked invasive and required neurosurgery, but what was allowed in a PG-13 movie. There were also just lots of lines of dialogue and phrases that the scientists use that we helped with.
SSF: The neuro-tech part of the film is reliant on several fairly specious scientific assumptions. Like, for instance, how close are we to actually uploading a monkey’s brain?
JC: The question is not how close we are to uploading a monkey’s brain, because I don’t know if that’s possible. We’d first have to spend hours getting into the philosophical argument about what is a mind. But what’s possible today and what our lab and Michel’s lab are doing is recording neuro-activity from the brain, which could be in a rat or a monkey or even humans. We can basically use machine learning and A.I. techniques to translate the activity of these neurons into some kind of intention, whether moving an arm or a computer cursor. That’s mainstream B.M.I. [Brain Machine Interface]—it’s a field that is happening today. But there’s a huge leap between that and talking about uploading subjective experience, or consciousness.
MM: I think there’s a large hurdle in the technology gap. Jose and I were coauthors of a paper, mostly driven by George Churchill’s group at Harvard, which asked whether you could physically record every neuron of a mouse brain. And the answer is that there is a fairly large gap between what we can do and what we need to do.
SSF: In your research, Michel, I read you developed a cyborg beetle. Can you talk about creating the computer brain for the beetle? And what is the difference between that entity and a sentient creature?
MM: If you rule out any ineffable stuff—call it “the soul”—and if you believe that the brain is purely mechanistic, then it just becomes a question of this: Is the machine we’re running [in our brain] at the pinnacle, and is the cellular-based fabrication process used to make it optimal? Or is it different because it’s on a different computational substrate, or is it different because it’s uses different algorithms, or is it different because it approaches the same problem from a different angle? And the answer is that we can’t answer that yet. Because we don’t have any instance of sentience not at the level of humans, and we just don’t know enough about the brain to approach it. Then again, nature didn’t invent radio frequency communication or turbines, so obviously there’s room within the universe for things that were not arrived at by the fabrication technology that we are made of.
SSF: Can you talk about the use of nanotechnology in the film? And how far-fetched is it in relation to what we have now?
MM: When people say nanotechnology in the entertainment business, they mean it as a catchall term for anything that looks cool and does something in a distributive fashion, but that term doesn’t do justice to whole fields. There are large disciplines and subfields: On the one hand, you have chemistry and chemically derived approaches, which makes sense because chemists have been working on that scale for a long time; on the other hand, you have things coming out of mechanical and electrical engineering, which work their way down and meet the chemists at the bottom, and the reality is that it’s a happy mixture of all of these things. Nanotechnology is everywhere these days, from the paint you buy at Home Depot to potato chips. Having said that, if you look at the approaches that people are taking to interface with cells at that scale, you get some interesting stuff: There’s been work recently exploring the use of nano-patterned substrates to record cells or neurons, or approaches to build tissue for regeneration. So we are building things at that level, but we’re obviously far from what’s in the movie. And those depictions violate tons of laws of thermodynamics.
SSF: What are you specifically referring to?
MM: For example, if you were to calculate how much energy and power would be required to reconstruct those solar cells in the movie at the rate that it happens, you would find some arguments. The rates that things happen for entertainment and dramatic reasons are probably a little over the top.
SSF: The representation of science in film often deals with the threat of technology. Transcendence seems to suggest this stuff is pretty dangerous.
MM: I disagree, not that it’s not dangerous, but that it makes scientists look bad. I don’t think it’s alarmist. I think any significant technological advancement comes with an ethical quandary. So you’ve got to get people thinking about that. And yeah, this one is far-fetched, but in terms of getting people to think about the issues, the movie is fairly benign, because in the end the machine wasn’t so bad. It doesn’t turn out to be Skynet.
SSF: When people discuss A.I., there seem to be two camps: those, like Ray Kurzweil, who say that the Singularity will be beneficent and constructive, and others, who say it’s destructive, like The Terminator’s Skynet. What do you guys think?
JC: There’s just not enough known. Sci-fi always asks this question: If you put consciousness or self-awareness into a machine, what’s going to happen? We just don’t know. The ancient philosophers were asking this same question, and there hasn’t been any progress in that sense. So I don’t worry about this.
SSF: Is there anything that particularly peeves your scientific brain about the movie?
MM: I honestly like the movie. But I did think it was amusing when the computer uploads itself to the satellite when the bad guys are approaching the abandoned warehouse. The way they finally show it makes it seem like this vast and amazing sentient machine uploaded itself to a satellite link in thirty-two seconds. And I can’t download a YouTube clip in thirty-two seconds. I thought it was funny. But it’s fine. I don’t think there’s anything worthy of kicking at.
SSF: What about the fact that this super-powerful being can’t keep the TV monitors from flickering?
MM: We talked about this. It was a stylistic thing. There were artistic choices made. There’s a limit when you approximate reality in these things, beyond which it stops being fun and becomes a university lecture.
SSF: What are the specific areas you are excited about in your research?
MM: “Neural dust” was this idea we published a white paper on in the summer, which asks the question of whether you can embed very small cell-diameter-size sensors in the brain cortex and record from them tetherlessly. And we are working hard on making it a reality.