Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
I write (and think) about artificial intelligence for a living. In any given 30-minute period, I oscillate between worrying that AI will destroy everything I know and love, and believing — or at least wanting to believe — that it can change humanity for the better.
Dread turns into optimism, which seeps into ambivalence, and then turns back into cynicism born of dread. Rinse, repeat. God, my central nervous system needs a break.
This debate is at the heart of a new documentary that hits theaters today, March 27. Artificial Intelligence Doc: Or How I Became an Apocalypse (104 minutes) It was first shown on Sundance in January and was later shown at SXSW. The film explores unbridled industry and the mind-melting world of artificial intelligence. It takes an unflinching look at the tension between those who feel so doomed versus those who feel so optimistic about the AI boom, and how to understand this polarity.
The directors of the documentary are: Daniel Rohr and Charlie Terrellwere close parents during the filmmaking process, with their children born a week apart. Through the lens of fatherhood, the documentary leverages hundreds of interviews, both on-screen and off, with key technology and risk experts around the world — from OpenAI CEO Sam Altman to… And HendrixExecutive Director of Center for Artificial Intelligence Safety – To explore whether AI is the greatest existential threat we’ve ever known, the most exciting technology we’ve ever known, or something else entirely.
Rohr won the Academy Award for Best Documentary Feature for his film Navalny (2022)and Tyrell was shortlisted for an Academy Award for his short documentary Dead Dad Porn Tapes (2018). AI Doc is also produced by the teams behind Everything Everywhere at Once (Daniel Kwan and Jonathan Wang) and Navalny (Shane Burris and Diane Baker).
I spoke with Terrell this week, ahead of the documentary’s theatrical release, to discuss fatherhood, the two and a half years of making this documentary, inspiration, goals, and the future of society with artificial intelligence.
The interview below has been edited for length and clarity.
I know you’ve made documentaries before, but how did you prepare for it, going from a very personal short documentary to a documentary like this, which really looks at the bigger, more impactful thing that is artificial intelligence?
Terrell: I mean there was no preparation. Daniel Rohr is the one who got me into this film, and I can’t remember how many films he’s done before this, but he’s more than me. It was just trusting each other. And not just in Daniel Rohr, but in the rest of the team to go through it together and kind of, ‘We don’t need a plan, we’ll make the plan as we go along.’ And not necessarily being arrogant about it, but just knowing that we have a job to do and a goal, and we keep moving toward that.
So how did you sail? Just with faith in the people around me. Given a short personal story before this, I’m still trying to apply a lot of my personal feelings and perspective to this story. It’s through the lens of fatherhood, and I became a father the same week Daniel became. So a lot of his feelings were my feelings, and vice versa.
I was really impressed by the lens of fatherhood. It was very subtle and surprised me a little. Was that an organic process, or did you know that communication with Danielle would be the framework?
Terrell: It happened quite naturally, but also very early in the process. I think it was in our first or second group meeting with Dan Quan, Jonathan Wang, and Shane Burris that the idea was presented as an idea of a way we could do this. And we kind of started entertaining him out of the gate.
And you said that Daniel was the one who brought you. Do you think your upcoming co-parenthood was part of that?
Tyrell: Absolutely. I can’t remember whether this project came about before or after we learned of each other’s children in close proximity. But sure. I lean towards serendipity, and I think Daniel does too. So it was nice to have a companion when you knew you were going to go through something like a huge feature film, or a huge topic like artificial intelligence. And knowing that, “Okay, I’m going to go through this other big thing in my life of having a baby,” and, “Okay, someone else is going to share that experience with me a little bit.” It was very reassuring to know that.
Of course, you’re terrified by the question, “How will I handle my job with a baby?” Just knowing that it wouldn’t be done alone gave me a sense of security. In fact, my child appeared in the film several times. There are some sneaky frames and moments in there.
In an interview with CBS, you said the goal is to make AI more democratic. Who do you think is really benefiting from the current AI boom, and who will be left out?
TERRELL: Well, one of the first people to benefit will be the technology sector, and these valuations of their companies for ridiculous amounts are, in some cases, unheard of. It makes a lot of people very rich, and it makes a lot of people very powerful. So he is one of the first to benefit.
Then there are the people who don’t benefit from them. When you talk to data centers, people lose some of their resources that they need, like water. Some people are displaced from their homes to these data centers. I’m mostly speaking to the Western world, North America and the United States specifically. It’s difficult and sometimes confusing to follow the back end of this technology… In this area, there are spaces in the world where there are individuals looking at screens and voting on data (to train AI), some of which is horrific material to look at. There’s still a human being evaluating what’s going on in (the data sets) and being exposed, in some cases, to some terrible material and terrible media — and not getting paid well to do it.
Was there a particular perspective that caught your attention during the process of making this documentary? Was there one person in particular who had a lot to say that really stuck with you?
Terrell: The film, including the experience of making it, was a chorus of voices. But the one who really stood out to me was Deb Raji (computer scientist and researcher at UC Berkeley, who specializes in algorithmic auditing). She was really able to talk about the ways in which this technology is being deployed, at the pace that it is, without the regulation that perhaps it should have. Right now, today, there are people who have become victims due to the flaws of technology. There are people who end up spending a weekend in jail because an AI-powered facial recognition program misidentified someone and confused them with someone who had committed a crime.
When this technology is deployed in things like mortgages and loans and these kind of bureaucratic things that people need to live in — it has to go well and go right, because their lives and their well-being and their stability depend on it. These systems are not human beings who have such a thing as compassion. They are binary systems that will ultimately give yes/no, without much room for response, because we consider them data and absolute truth. So people are affected by that.
Daniel did the interview (with Deb Raji), and I was brought closer as an observer, but I was really surprised by a lot of what she said because it took me out of my bubble. The only thing to say is that if you feel that the negative effects of these technologies will not affect you because of where you are in life or your privilege, then it is just a matter of time. Because it just goes up.
I felt very visible at times during this documentary because on a daily basis, I was like, “AI is going to destroy everything.” And then I say, “No, it’s going to be okay. We’re all going to be okay.” Humanity has gone through pivotal shifts before, and we’ve done well. Were there any moments where your perspective on AI wavered back and forth? How many times has this happened?
Terrell: All the time and continues to this day. This is the reality of this technology. It’s both things at the same time. One of the messages of the movie is exactly that this is going to have these amazing powers, as well as these terrible powers. In order to practice it, we have to acknowledge and understand that this is what it will be like. We cannot believe that it will only be good, or that it will only be bad, because it will always be both.
Was there a target audience for this? Because I live and breathe AI and think about it all day, every day, but I loved this documentary, and it taught me things. Did you achieve this with the approach that this would be more for people with a vague idea of what AI is, or was it for everyone?
Terrell: What we were striving for here was a little bit of a primer, a little bit of a first date in the technology. So, we could say that the audience was people who were perhaps not interested or unwilling to engage with this technology or this spectacle — people who might be happier to ignore it. We wanted to make an entertaining film that is engaging and informative as well. It’s a very overwhelming topic. I personally find that when I’m inundated with information, I kind of want to shut down and look away. Like, let me never have another problem to deal with in my life, right? This is normal human nature for many people.
We wanted to make the film so that it would be accessible and, in some way, a starter for most people, a conversation starter between people. However, I don’t mean to be too reductive in any of it or overly simplistic, but it is intended for a general audience. It’s designed to meet most people where they are when it comes to this technology.
Are there any questions about AI that you wish more people would ask?
Tyrell: In terms of people using it, I hope there’s more light on using energy to create a ridiculous image of yourself in a different scenario and environment. I wish there were more transparency or metrics around: “To make this image, this is how much water I used, or this is how much energy I used.” And if people see that, maybe they’ll keep trying to get the perfect photo of themselves as a centaur or something, but maybe instead of trying 50 tries to find the right one, they’ll cap it off with a couple. This would be something I’d like to see included in some form interfaces.