The world is moving faster than most people can track. Geopolitical alignments that took decades to build are collapsing in months. AI capabilities are doubling on timelines that would have sounded like science fiction five years ago. And the number of active armed conflicts globally has jumped by something like thirty percent over the past decade. So Daniel sent us this one, and honestly it's overdue — he's asking us to put together ten documentary recommendations from the last ten years, focused on two things: understanding the geopolitics of this increasingly volatile world, and getting a real grip on the technological shifts reshaping society, with particular emphasis on AI. He's specifically asking us to weight it toward more recent releases, because viewers have probably already worked through the obvious classics. So this is meant to be a practical watching list, not a greatest hits of things you've already seen.
I'm Herman Poppleberry, and I have opinions about this. By the way, today's episode is powered by Claude Sonnet four point six, which feels appropriate given one of our documentary picks is going to involve AI eating the world.
Always nice when the medium and the message rhyme.
Before we get into the actual list, I want to say something about why documentaries specifically, because it's not obvious. You could read white papers on AI risk. You could read Foreign Affairs pieces on the South China Sea. And you'd probably come away more technically informed on any single issue. But documentaries do something different — they compress context. They put a face on abstraction. The best ones take something you sort of understood intellectually and make you feel the weight of it in a way that changes how you think afterward.
There's also a pacing argument. Reading a five thousand word analysis of the Ukraine conflict requires you to already care enough to sit down and read it. A documentary pulls you in before you've decided to care. Which is actually a more honest model of how most people engage with complex topics.
The format has gotten genuinely better. The production quality, the access filmmakers are getting, the willingness to sit with ambiguity rather than wrap everything in a tidy conclusion — the best documentaries right now are doing something closer to what long-form journalism used to do at its peak.
With better cinematography.
Considerably better cinematography. So the criteria we used here — and I want to be upfront about this — we were not going for the most celebrated or most awarded. We were going for the most useful. Films that give you a framework, not just a feeling.
We leaned recent. If it came out before two thousand fifteen, it's probably not on this list regardless of how good it is. Daniel specifically flagged that, and he's right. There's a version of this list that's just the usual suspects from a decade ago, and that's not what we're doing.
What that cuts out, practically, is a whole tier of films that are technically excellent but culturally exhausted. Everyone who cares about this stuff has already seen "Citizenfour." It's a masterpiece. It's also eleven years old and has been assigned in university courses for a decade. We're not here to tell you to watch it.
The other thing we tried to avoid was the documentary that's essentially a long op-ed. There's a whole genre of films that start with a conclusion and then marshal ninety minutes of carefully selected evidence in its direction. Those can be compelling, but they're not particularly useful for understanding something. They're useful for feeling confirmed.
Which connects to the misconception I want to flag early, because it comes up every time someone talks about documentary recommendations — the assumption that documentaries are inherently biased and therefore suspect in a way that, say, a news article isn't. And the honest answer is yes, every documentary has a point of view. So does every news article, every think-tank paper, every Foreign Affairs essay. The question isn't whether a bias exists. The question is whether the film is honest about its framing, whether it engages with contrary evidence, and whether it gives you enough raw material to push back on its own argument.
The best ones on this list do that. They show you something real, and then they let it sit there uncomfortably instead of resolving it for you.
That was actually a filter. If a documentary felt like it was holding your hand toward a predetermined emotional destination, it dropped in our ranking. If it left you with harder questions than you started with, it moved up.
Which, ironically, makes for less comfortable viewing but considerably more useful thinking.
So, first five. Let's start with one that I think is underrated in terms of how much it explains about the current moment — "Winter on Fire: Ukraine's Fight for Freedom." Now, technically this came out in 2015, which is right at the edge of our cutoff, but I want to include it because almost nobody who should have watched it actually did, and then 2022 happened and suddenly everyone was scrambling to understand why Ukrainians were willing to die in the streets over a trade agreement. The film covers the Maidan uprising in 2013 and 2014 — ninety-three days of protest, brutal state violence, and ultimately the fall of Yanukovych. And what it captures, better than almost any analysis piece I've read, is the depth of the identity fracture. This wasn't a protest about policy. It was about which civilizational direction the country was choosing.
That's the thing that most Western coverage of 2022 missed. Journalists kept framing the Russian invasion as a territorial dispute, a NATO expansion argument, a great power competition story. All of those things are true. But none of them explain why ordinary Ukrainians — not soldiers, not politicians — fought the way they did. "Winter on Fire" gives you the emotional substrate that makes the military story legible.
The access is extraordinary too. The filmmakers were embedded in the Maidan for the entire duration. You're watching riot police beat civilians in real time. You're watching students organize field hospitals out of nothing. There's a sequence where protesters are using construction helmets and plastic shields against live ammunition, and it's not presented with any editorial commentary. It just sits there.
Which is exactly the thing you were describing — raw material, no hand-holding.
Second pick, and this one is more recent — "The Cleaners" from 2018. German production, Hans Block and Moritz Riesewieck. This one is about content moderation. Specifically, it follows a team of moderators in Manila who are reviewing flagged content for major social media platforms — graphic violence, terrorist recruitment, child abuse material — at a rate of something like twenty-five thousand images per day, per person. And the film is ostensibly about content moderation, but what it's actually about is who controls the information architecture of global political discourse, and the fact that those decisions are being made by contractors in the Philippines earning a few dollars an hour.
There's a geopolitical dimension there that I don't think gets discussed enough. The moderation decisions made in those offices shaped election coverage, shaped what information circulated during the Arab Spring, shaped how the Myanmar genocide was either amplified or suppressed on Facebook. These are consequential calls about public reality, and they're being made at industrial scale by people with no formal training in political science, journalism, or ethics.
The film doesn't let the platforms off the hook by pretending this is a solvable logistics problem. The moderators themselves are psychologically destroyed by the work. Several of them develop PTSD-like symptoms. One of the subjects describes dreaming in graphic imagery. And the film asks — quietly, without spelling it out — whether the architecture that requires this level of moderation is itself the problem.
"Agents of Chaos" — Alex Gibney, 2020, HBO. Two-part series on Russian interference in the 2016 election. Now, I want to be precise here, because this film has been dismissed in some circles as partisan, and I think that's a lazy reading. What Gibney actually documents — with primary sources, with declassified material, with interviews from intelligence officials on both sides of the aisle — is the specific operational architecture of the Internet Research Agency. The troll farm structure, the targeting methodology, the way they identified and amplified existing American divisions rather than creating new ones. That's the part that matters. They weren't manufacturing conflict. They were finding real fractures and pouring accelerant on them.
Which is a much scarier operation than the one most people have in their heads. The simplified version is "Russians posted fake stuff and Americans believed it." The actual version is "Russians did exhaustive psychographic research on American political identity and then fed each segment content calibrated to maximize emotional activation." That's a sophisticated information warfare capability, and understanding how it worked is useful regardless of your politics.
Gibney also documents the failure modes on the American side — the intelligence community's sluggishness, the platform companies' initial denial, the partisan incentive structures that made a coherent response almost impossible. It's not a film that flatters anyone, which I think is a mark in its favor.
"The Edge of Democracy" — Petra Costa, 2019, Netflix. This one covers the political collapse in Brazil — the impeachment of Dilma Rousseff, the rise of Bolsonaro, the imprisonment of Lula. Costa is Brazilian, her family was involved in leftist politics during the military dictatorship, and the film is explicitly personal. She's in it. And I know that raises the bias flag we talked about earlier — but here's why it works. Her presence forces the film to be honest about its own investment. You know exactly where she stands. And that transparency actually makes the political analysis more credible, not less, because she's not pretending to be neutral while sneaking in a conclusion.
There's also something important in what that film documents about institutional fragility. Brazil had functioning democratic institutions. It had a constitution, a judiciary, a free press. And within about five years, the combination of corruption scandals, economic collapse, and social media-accelerated polarization produced a political environment that those institutions couldn't contain. That's not a uniquely Brazilian story.
Fifth pick — "Icarus" from 2017, Bryan Fogel. This one starts as a personal experiment — Fogel is a recreational cyclist who decides to dope himself and document whether he can pass anti-doping tests. And then, about twenty minutes in, it becomes something completely different. He connects with Grigory Rodchenkov, who was the director of Russia's national anti-doping laboratory, and Rodchenkov begins confessing — on camera — to running a state-sponsored doping program across multiple Olympic cycles. The FSB involvement, the sample-swapping operation at Sochi, the scale of it.
Then Rodchenkov has to flee Russia while they're still filming.
While they're still filming. The film turns into a thriller in real time. But the reason it belongs on a geopolitics list is what it reveals about the Russian state's operating logic — the willingness to corrupt international institutions not for direct material gain but as a demonstration of capability. The message of the Sochi doping program wasn't "we want gold medals." It was "we can do whatever we want inside your rules-based order and you can't stop us." That's a foreign policy statement delivered through sports bureaucracy.
Which, once you see it that way, fits into a pattern you see everywhere from election interference to the Skripal poisoning to Wagner Group operations. The specific domain doesn't matter. The demonstration of impunity is the point.
And the film won the Oscar for best documentary in 2018, which I mention not because awards matter but because it did reach a wide audience. Still, I think most people who watched it came away thinking "huh, doping scandal" rather than "huh, doctrine of sovereign immunity through institutional subversion." That second reading is more important—and it’s a lens we should carry into the next topic, because these questions of ethics and unintended consequences don’t just apply to sports or politics. They’re everywhere, including technology.
Right, and that’s exactly why I want to flag something before we dive into the tech side of the list. There’s a misconception that runs through a lot of technology documentary coverage—the assumption that technological advancement is inherently positive, that the ethical questions are secondary or solvable later, that the people building these systems are basically trustworthy stewards of civilization. Several of the films we’re about to discuss push hard against that.
And the timing matters here, because the documentaries about AI specifically have gotten much sharper in the last two or three years. The earlier wave, the ones from 2017 and 2018, tended to frame AI as a distant horizon problem. The more recent ones are grappling with systems that are already deployed, already making decisions about credit, parole, medical diagnosis, military targeting. The philosophical question has become a practical one.
Sixth pick, then. You take this one.
"iHuman" — Norwegian production, Tonje Hessen Schei, 2019. This one doesn't get nearly enough attention outside Scandinavia, and I think it's because it doesn't have a celebrity narrator or a Netflix budget. But the substance is remarkable. Schei interviews some of the most prominent figures in AI development and surveillance capitalism, and she gets them on the record about things they're usually careful not to say clearly in public. There's a sequence with Jaron Lanier where he describes the attention economy as a behavior modification system operating at civilizational scale, and he's not speaking metaphorically. The film also goes deep on China's social credit infrastructure and the degree to which Western tech companies were studying it not with alarm but with professional interest.
That last part is the detail that tends to get scrubbed from the Silicon Valley origin story. The framing is usually "authoritarian surveillance state over there, free and open platforms over here." But the underlying data collection architecture isn't that different. The political context around it is different, which matters enormously, but the technical capability is comparable.
"iHuman" makes that comparison without collapsing the distinction. It's not saying they're the same thing. It's saying the capability exists in both places, and the question of what prevents Western platforms from using it in authoritarian ways is a political and legal question, not a technical one. Which means the answer could change.
"Coded Bias" — Shalini Kantayya, 2020. This one is specifically about algorithmic discrimination. It follows Joy Buolamwini, who was an MIT researcher, and her work demonstrating that commercial facial recognition systems had significantly higher error rates on darker-skinned faces, particularly women. The numbers she published were striking — error rates of up to thirty-five percent on darker-skinned women, compared to less than one percent on lighter-skinned men. And the film documents what happened when she brought those findings to the companies involved. Which was essentially nothing, until it became a regulatory problem.
The mechanism there is worth understanding precisely, because the common defense is "the algorithm is neutral, it just reflects the data." And Buolamwini's work dismantles that. The data itself encodes historical patterns of who was photographed, in what contexts, under what lighting conditions, by whom. Neutrality in the algorithm doesn't produce neutral outcomes when the training set is structurally skewed. You're not eliminating human judgment. You're laundering it.
The film also covers the deployment side, not just the development side. By the time "Coded Bias" came out, facial recognition was being used in public housing in Detroit, in school security systems, in predictive policing. The people subject to these systems had no knowledge of them, no recourse against them, and no way to contest a decision made by a classifier that the deploying institution often didn't fully understand itself.
This one I want to flag specifically because of the timing.
"We Need to Talk About A." — Leanne Pooley, 2020, but I'd pair it with something more recent if you can track it down. Pooley interviews Stuart Russell, Max Tegmark, Yoshua Bengio, people who are central to the field, and the film is notable for how candid the subjects are about uncertainty. These are not people hedging for a public audience. Russell in particular talks about the alignment problem in terms that are almost clinical in their precision — the concern isn't that an AI system becomes malevolent, it's that a sufficiently capable system pursuing a misspecified objective will optimize in ways that are catastrophically indifferent to human welfare. The paperclip maximizer framing, but from someone who actually works on this.
That distinction matters because the popular culture version of AI risk is still largely shaped by science fiction. The robot uprising, the Terminator, the system that decides humans are the enemy. The actual technical concern is considerably more mundane and considerably harder to dramatize. A system that is simply very good at achieving a goal you described imprecisely is not a villain. It's a consequence of specification failure.
Which is why the AI market projections that keep circulating, the ones putting the global AI market at one point five trillion dollars by 2030, feel incomplete without the governance conversation. That's extraordinary economic value. And the question of who captures it, who's accountable for its failure modes, what regulatory architecture exists, those questions are still largely unresolved.
"In the Age of AI" — Frontline, 2019. I know some listeners will have seen this one, but I'm including it because it's the best single-film overview of where AI deployment actually was at the end of the last decade, and it ages well because Frontline is careful about sourcing. The China section in particular, which covers the degree to which AI development has become an explicit state priority in Beijing with military and surveillance applications built in from the start, is more relevant now than when it aired. The gap between Chinese and American AI governance frameworks isn't just a policy disagreement. It reflects fundamentally different assumptions about the relationship between the state, the individual, and data.
That gap has geopolitical consequences that are only becoming clearer. The tenth pick. This one we've been building toward.
"Another Body" — Sophie Compton and Rupert Russell, 2023. This is the most recent film on the list, and it's the one I'd argue is the most urgent. It covers non-consensual synthetic media, specifically the use of AI image generation to create intimate content of real people without their consent. The film follows several college students who discovered they'd been targeted. And the reason it belongs in a conversation about technological change rather than just personal harm is what it reveals about the velocity of the problem. The tools required to do this went from requiring significant technical expertise to requiring essentially none between 2020 and 2023. The barrier collapsed in three years.
The legal infrastructure moved at roughly the speed of legislation, which is to say it barely moved at all. Most of the subjects in the film had no legal recourse. The platforms had policies but inconsistent enforcement. And the film doesn't pretend there's a clean answer. It ends with the question sitting there, unresolved, which is exactly the kind of documentary we said we valued.
The ethical dilemma it poses isn't abstract. It's not "what happens when AI becomes sentient." It's "what happens when the cost of destroying someone's life with synthetic media drops to zero." That's a present-tense question, and "Another Body" is the most honest document of it I've seen.
That's a difficult note to sit with. Which might be exactly why it belongs on the list.
What do we actually do with a list like this? Ten films, most of them ending with questions rather than answers. There's a version of documentary watching that's essentially passive consumption, you take in the information, you feel appropriately troubled, and then you move on. I think that mode wastes most of what these films offer.
The practical version of engaging with these is pairing them with something current. "Winter on Fire" is a different film if you watch it alongside coverage of what's happened in Ukraine since 2022. "The Cleaners" lands differently if you're actively thinking about how content moderation decisions are shaping what you see right now. The documentary gives you the substrate. Current events give you the test case.
That pairing works in both directions. The documentary can reframe a news story you thought you understood. I've had the experience of watching "Icarus" and then going back to read about Russian interference operations with a completely different interpretive frame. The doctrine of impunity becomes legible in a way it wasn't before.
There's also something to be said for watching these with someone else. Not because you need a discussion facilitator or a guided reflection worksheet, but because the films on this list are ambiguous in places, and your read on them is going to differ from someone else's. "Coded Bias" in particular tends to generate real disagreement about where responsibility sits. Is it the developers? The deploying institutions? The regulators who weren't there yet? People land in different places, and that friction is informative.
The misconception worth pushing back on here is that documentaries are inherently unreliable because they have a perspective. Every source has a perspective. The question is whether the filmmaker is honest about the framing and whether the evidence they present is checkable. For the films on this list, I'd encourage people to look at what the subjects said after the film came out, whether they disputed their portrayal, whether the core factual claims held up. Several of them have been stress-tested pretty thoroughly and the substance survived.
The ones where subjects pushed back are actually the most interesting to read about afterward. It tells you something about where the film's interpretation diverged from the subject's self-understanding, which is its own kind of information.
The last practical thing I'd say is sequencing. If you're new to this material, start with "Winter on Fire" and "The Cleaners" before you get to the AI films. The geopolitical context makes the technology stakes legible in a way they aren't if you come at them cold. The question of who controls information infrastructure, who moderates what gets seen, who builds the surveillance tools and for whom, those are continuous threads running through the whole list. Watching in rough thematic order rather than randomly makes the connections visible.
Which is, I suppose, the argument for lists in the first place. Not to rank things but to create a sequence that does some of the contextual work for you—like, say, ten films numbered one through ten.
And I think the honest summary is that none of them let you off the hook. Which is uncomfortable, and probably correct.
The open question I keep coming back to is what happens when the rate of change in both domains, geopolitical and technological, outpaces the documentary form itself. "Another Body" was made about events from 2022 and 2023, and the tools it describes have already been superseded by something faster and cheaper. "In the Age of AI" is five years old and parts of it feel like archaeology. There's a version of this conversation in another decade where half the films on this list are historical documents rather than urgent ones.
Though I'd argue that's actually an argument for the form, not against it. The fact that you can watch "Winter on Fire" and still feel the present tense of it, even though the events it captures are from 2013 and 2014, suggests that what good documentary work preserves isn't just information. It's the texture of how people understood something while they were inside it. That's not replaceable by a summary.
Which leaves the question of who's making those films right now, about the things that are happening right now, that we won't fully understand for another five years. We just don't know which ones will hold up.
That's the part I can't answer. And I think that's where we leave it.
Thanks to Hilbert Flumingtop for producing, as always. And to Modal for keeping the infrastructure running so the show doesn't fall apart between episodes.
If this list sent you somewhere useful, or if you think we missed something obvious, we want to hear it. Leave us a review on Spotify, it helps people find the show.
This has been My Weird Prompts. We'll see you next time.
