The immediate response to Darren Aronofsky’s new, AI-generated drama made me think of Cersei Lannister’s public flagellation in Game of Thrones. The great director and long-time provocateur was Lannister, walking naked through the town, the online commentariat pelting him with fruit, everyone yelling “shame!” The Guardian called his shortform series, titled On This Day…1776, “a horror.” Gizmodo said it “looks like dogshit.” According to Kotaku, it’s “as bad as it sounds.” Not to be outdone, CNET declared that Aronofsky is “ruining American history.”
Releasing an AI-generated scripted production, at this point in time, is a ballsy move for a marquee filmmaker, given Tinseltown hasn’t exactly embraced the technology with open arms. Not because it’s a flash-in-the-pan innovation, destined to go the way of the dodo. Quite the opposite: the industry clearly believes the writing on the wall says “this tech is here to stay, and it’s only going to get better.”
Aronofsky executive produced On This Day…1776, a series of weekly videos landing on TIME’s YouTube channel, depicting various moments from America’s Revolutionary War. In the first episode, King George III announces to Parliament his plans to “put a speedy end” to efforts to establish an independent empire, deploying battleships and soldiers aplenty, which doesn’t exactly calm things down. In the second, Benjamin Franklin recruits Thomas Paine to write propaganda advocating for “the cause of America.”
If you know anything about AI-generated video, you’ll know it looked embarrassingly bad just a few years ago. The improvement has been astronomical. What might it look like in two years? Five? Fifty? Take the long view, and On This Day easily resembles a historical curio, comparable to splotchy short films like The Kiss and The Sprinkler Sprinkled—two of the earliest motion pictures, released in the late 1800s. Nobody considered them great works of art; they matter because they were created by people who—like Aronofsky—came to the party early.
These are the people who do indeed tend to get pelted with fruit. Cinema was initially dismissed as too mechanical to qualify as “real” art; photography was announced as the death of painting; television would supposedly end human conversation. I’m not suggesting that contemporary anxieties about AI-generated video are unfounded. Only that they sit within a historical continuum involving cycles of hostility, often taking the form of moral panic, which greet emerging forms before the dust settles and the medium in question develops its own grammar, syntax, audience, etcetera.
Primordial Soup via YouTube/Screenshot by Flicks
One Aronofsky fruit-pelter—Angie Han of The Hollywood Reporter—clearly wanted to be the septa behind the director, bell in hand, presiding over the scolding. In a venomous take-down, she even had the gall to argue that Aronofsky’s decision to embrace resource-intensive AI makes him a hypocrite when it comes to his environmental activism (because, don’t you know, eco-activism demands monastic purity: one must renounce all devices and clothe themselves in hand-stitched bamboo). This breathtaking pettiness reminded me of that old admonition: play the ball, not the person. The stuff to criticise is what you see on the screen; the work itself.
Han made a better point when she described the series as having a “plasticky sheen.” That’s true, though the same criticism could be made of every MCU movie: this is just a different kind of polish. In any event, the point distracts from the production’s bigger problems: the script, which is larded with inert dialogue, and its erratic structure, which jumps around without ever properly grounding the viewer in interesting drama. Part of this comes from a determination to keep the pace brisk, and part from an impulse to show off the technology—shuffling people and locations with restless speed.
Going by the first two episodes, the series is at its best when its visual inventions serve the story—narrative, technology, and momentum moving in sync. About a minute into the first episode, for instance, there’s a brief montage depicting a wax-sealed royal decree being passed from hand to hand.
Primordial Soup via YouTube/Screenshot by Flicks
Each shot maintains the same compositional structure, but the figures change as the document travels from Parliament to the British military headquarters in Boston. The montage lasts barely 10 seconds, but it’s a compact example of bold visual storytelling. And it’s the sort of flourish that AI is well suited to, generating variations on a single image.
This points to an interesting question, largely ignored by the heckling peanut gallery: what, narratively speaking, are the strengths and weaknesses of AI-generated video? What does it excel at, or show great potential for, and what kinds of storytelling techniques does it struggle with or undermine? Perhaps it’s too early to answer these questions with any confidence. But given the technology isn’t going anywhere—we can all agree on that—it’s worth beginning the conversation now, while the grammar of the medium is still being written.
