TikTok Cinema: Inside AI Microdramas

A content machine

Thefirst wave of AI-generated videos was largely the territory of tech bros experimenting with the limits of technology, but as generative video models become more sophisticated, agrowing industry has been emerging around turning those experiments into commercially viable entertainment. The emphasis seems to be less on telling better stories than producing more of them. And the industry isn’t limiting itself to new material, especially when there are opportunities to capitalise on the enduring commercial value of celebrity. Case in point: theAustralian soapNeighboursis being recut into AI-assisted microdramas turning Margot Robbie’s old storylines into anew product for our phone screens, exploiting and extracting yet another product from afamiliar face. For atechnology repeatedly promising to​‘redefine storytelling’ and unleash human imagination, it’s aremarkably unimaginative ambition.

An AI-generated series can currently be produced in just two weeks, compared with 12 for live-action. The most popular apps offer creators anAI production toolkit complete with automated storyboarding, visual sets, voice cloning and dubbing in more than 40 languages. Shortical, amicrodrama app rapidly expanding into AI (PSA for the discerning: it’sIsraeli-owned), claims its ten highest-earning creators made over$2 million in asingle quarter, and promises to turn anidea into aglobal hit within weeks. For ageneration facing increasingly precarious creative careers and questioning whether their work will ever matter, the possibility of making aliving this way is hardly insignificant.

As the industry celebrates how few people it takes to make aseries today,actors are already losing roles to people who don’t exist, while their replacements appear to have inherited some remarkably familiar beauty standards. The aesthetic that social media spent years encouraging real people to aspire to – the eerily uniform, racially ambiguous, but Western-centred ideal of beauty known asInstagram Face – can now be generated without them. From my own extensive – if regrettable – viewing experience, AI microdramas seem particularly fond of alimited gene pool of doe-eyed women and vaguely Mediterranean-looking men, seemingly plucked from aLove Islandcasting call. Atechnology capable of conjuring almost anyone, determined to reproduce the same handful offaces.

Even the peculiar world of celebrity worship is finding ways to dispense with real people. One AI microdrama that found fame through TikTok has amassed over 120M views, with its creator setting up social media profiles for the two leads. Between them, they’ve quickly attracted close to 100K followers, some of whom speak directly to the characters as though they were real. Meanwhile, the person behind them remains entirely invisible, referred to by fans simply as ​‘the creator’: arather grand title in an industry concerned with how little creating its creators need todo.

But there’s adifference between making filmmaking more accessible and making the production of content more efficient. These platforms aren’t just competing with TV, they’re competing with scrolling – during commutes, queues and chores, for every spare moment of our lives. It’s no surprise in this race to produce more, the distinction between having the tools to makeanything and having the time, freedom and imagination to create something worth watching is becoming blurred.

The hassles of storytelling

Storytelling has never been about arriving at the end of aplot. Long before they were written down, stories have passed between people, carrying their fears, desires, histories and ways of understanding the world. Irish storytellerClare Murphy speaks of how reflecting on our experiences allows us to turn them into stories, transforming what we’ve lived into knowledge and, when shared, wisdom. More than just asequence of events, astory is somebody’s attempt to make sense of it all. There are as many ways to tell stories as there are people telling them. Myths and romances have been borrowed and retold but each retelling can bring anew understanding of the world, shaped by the experiences of the person telling it.

AI, on the other hand, can reproduce the patterns of stories without having to make sense of the experiences that gave rise to them.In one rather bleak experiment, researchers asked an AI model to generate over 11,000 stories from 236 different countries. Despite the enormous cultural differences, it repeatedly produced variations of almost exactly the same plot. Remembering the contents of astory isn’t the same as understanding why someone needed to tell it in the firstplace.

This is more than atheoretical concern. Having watched afrankly concerning number of AI microdramas while researching, I’ve encountered almost identical scenes with little more than achange of setting. Writer and cultural criticKyle Chayka has written about how algorithms have spent years pushing creators towards ​‘flattened styles’ of similar work.As he puts it, the ​“fastest route through the AI machine is to get acliché: what AI produces is fundamentally kitsch”. Combine this with platforms measuring success in views, shares and watch time, and it’s hardly surprising that the results look so alike. The artistic ambition seems to be capped at whatever the algorithm already knows willsell.

A film (or series) however, is rarely the work of asingle imagination. The actors, writers and crews being displaced aren’t interchangeable parts of aproduction line. There is beauty in the number of minds that come together – their personalities, experiences and creative disagreements all shaping what eventually appears on screen. Animprovised performance can reveal something about acharacter the writer hadn’t previously considered. Better yet, human tensions can find their way into the work itself (lest we forgetDon’t Worry Darling). The finished product can become something nobody could have imagined at the outset. Regenerating an AI scene to correct distorted faces or disappearing limbs might require effort – not to mention considerable natural resources – but these glitches are problems of execution. The most valuable struggles of storytelling aren’t technical. They’re empty pages, abandoned ideas or moments of uncertainty when acreator has to question what they’re trying to say in the first place. It’s often through that process that something genuinely unexpected emerges, rather than another version of what already exists.

Hayao Miyazaki puts it best: ​“all important things in the world are ahassle”. An industry determined to eliminate that hassle loses the human experiences that make it worth watching.

The price of beingseen

China, where microdramas first took off, offersaglimpse of where this all might lead. In the first eight months of 2026 alone, some 430,000 microdramas were released, with more than 90% reporting some use of AI technologies. An extraordinary amount of entertainment, all competing for aplace in an already overcrowded attention economy.

As production becomes cheaper and content more abundant, the platforms controlling distribution become more powerful.Research into China’s microdrama industry found that filmmakers often had limited access to audience data, sometimes receiving little more than figures showing whether viewers were still watching three, five or ten seconds into an episode. Filmmakers have always had to satisfy the people financing their work, but they’re also making stories for systems that measure their worth in afew seconds.

There’s acruel irony to an industry promising to remove the financial barriers to filmmaking, only for the cost of getting anyone to actually watch your film to become the bigger obstacle. Without free promotional support, producers reported that getting adrama seen could meanspending ten times its production budget on marketing, effectively paying to push their episodes into more people’s feeds. In the US, microdrama companies are pouring advertising money into Meta’s platforms, with some dedicatingas much as 75% of their budgets to the tech giant. And through all this, Meta benefits twice: collecting advertising revenue whilst the promotional episodes themselves keep audiences scrolling on its platforms.

For all the promises of democratising filmmaking, the power to reach an audience remains concentrated in remarkably few hands. AI might allow more people to make films without major studio blessing, but it does little to free them from the demands of the platforms distributing their work. The tools might be available to everyone, but the audience certainly isn’t.

And what of the audience? Unlike your run-of-the-mill doomscroll where each video is forgotten almost as quickly as it’s watched, Ifound myself with agrowing collection of unfinished stories rattling around my head. Rather than arevolution in entertainment, AI microdramas are the inevitable conclusion of what the internet has been becoming for years: an endless procession of influencers, trends and cultural products competing for the same shrinking sliver of attention. Everything must be immediate and, more importantly, short enough to watch whilst waiting for the kettle toboil.

When every spare moment is swallowed by another, there’s less room for our own minds to breathe. As the industry races to manufacture ever more stories, Iwonder what happens to the ones we might have dreamed ourselves.