Hi there! My name is Janko Roettgers, and this is Lowpass. This week: AI is coming for microdramas.

For better or worse, AI microdramas are here 

Have you ever watched a microdrama, perhaps while scrolling through your TikTok feed, and thought: This all looks kind of fake?

Turns out you might be onto something – and it’s not just the constant cliffhangers, the overly dramatic expressions and the barely believable plot twists. Microdrama producers and publishers are increasingly looking to AI to create shows at scale for less, and AI companies are starting to discover microdramas as the next frontier for synthetic entertainment.

This summer has been awash with AI-related  microdrama announcements: Earlier this month, Israel-based microdrama app Shortical raised $100 million, thanks in part to an AI-generated show that features a shirtless price, an evil queen, and a bunch of dragons. In June, Beijing-based Storeel launched an AI production platform for microdrama creators.

Next week, Berlin-based Inkitt is scheduled to unveil Ironblood, a dedicated platform for AI-generated Sci-Fi and action microdrama “movies.” And today, Character.ai is adding a first slate of microdramas to its mobile app, with plans to merge traditional short-form storytelling with interactive AI chat.

Cheaper, faster … better?

When Hollywood executives talk about AI, they often stress that they don’t intend to replace human actors or storytellers. The goal of using AI tools is not to make stories cheaper or faster, but better, or so the mantra goes for fear of backlash from audiences and creatives alike.

Microdrama creators generally don’t have the same kind of qualms. Their shows are already produced for cheap, often feature little-known talent, and for the most part exist outside of the reach of Hollywood’s unions. Consequently, the companies that are looking to use AI are open about the potential cost savings that motivate them.

StoReel told Business Insider this spring that it can make an hourlong microdrama for as little as $20,000 with AI, while making the same show with human actors could cost up to ten times as much. TrueShort, a yet-to-launch AI microdrama app, told the same outlet in May that it could produce a 30-minute AI movie for as little as $3000.

The only question: What do you actually get for these kinds of budgets? Early examples suggest the results can be a bit of a mess. Take Shortical’s Bound By Fire AI microdrama, for instance. Prince Chris, the handsome slash gruff male warrior lead of the show, appears to wear a completely different wardrobe after walking from one room to another. 

His secret 8-year-old child is almost as tall as her mother in one shot, only to be a whole head shorter in the next. There’s also a plate of food that appears out of nowhere, a ponytail that magically switches from one shoulder to the other between shots, a dress that keeps changing its pattern, and a freestanding castle tower that doesn’t make any sense whatsoever. But at least the dragons look just like those of a popular TV franchise …

Viewers are starting to take notice

Some microdrama viewers are starting to push back against the increased use of AI in the genre. “I hate the [...] reality-bending nature of it all. If this continues to track, it’s going to kill vertical dramas for me,” one viewer recently noted on Reddit. “The AI is so bad and inconsistent,” another viewer noted. “Accents changing mid sentence, names being pronounced 15 different ways, faces looking completely different in different shots.”

“I’m so done with the AI ones,” a third viewer vented.

The only problem? Publishers may not care. Much of the microdrama boom is built around short-term growth, with startups raising tens of millions of dollars, only to spend them as fast as they can on paid social ads designed to lure in new viewers. Once people download these apps, they’re confronted with user-hostile business models designed to charge them huge amounts of money as quickly as possible, with subscription fees that make Netflix’s most expensive tier look like a steal in comparison.

For more on this, read my recent story Hollywood is in love with microdramas, for all the wrong reasons

That’s not how you build a sustainable business. It does however create the illusion of massive future windfalls, which has been enough to get VCs and Hollywood studios to open their check books. And if AI helps to lower production costs, then more of that money can be spent on user acquisition. That brings in more overpriced subscriptions, further fueling the illusion of a profitable industry, and thereby attracting more investor money. Rinse and repeat.

There may be another way

Given the current state of AI microdramas, it’s refreshing to see a company take a different path, and perhaps a bit surprising that the company in question happens to be Caracter.ai. Not only has Character.ai dealt with its own fair share of controversies in the past, it also has both the data and the user base to quickly churn out countless hastily prompted microdramas – but chose not to.

Instead, Character.ai is debuting just three animated microdramas today that it produced with help from Hollywood scriptwriters and a small in-house studio. “We're not trying to create a programmatic machinery to take any chat that happened [on Character.ai] and create a microdrama,” Character.ai CEO Karandeep Anand says. “That [would be] a recipe for AI slop. We are anti slop.”

WIth monetization through subscriptions and advertising already in place, Character.ai also doesn’t have to copy the exorbitant fees that are so prevalent on many microdrama apps. Anand tells me that microdrama monetization is a secondary concern for the time being. The bigger challenge: Will Character.ai users actually see these shows as something worth their time?

To win them over, the company is approaching microdramas as just one piece of a larger interactive entertainment ecosystem that also includes books and comics. “For us, microdrama is a very natural extension of the role play that's already happening on the platform,” Anand says.

Users are able to chat with the characters of each show, ask them questions about the plot, and even develop their own storylines to expand each show’s universe. “You jump into these stories,” Anand says. “You'll live these stories with these characters. We want you to jump in and chat with these characters and take the story forward.”

Another thing that’s refreshing about Character.ai’s approach? Instead of selling the world on vertical storytelling being the next big thing, Anand openly admits that it’s a bit of an experiment, and that he ultimately doesn’t know how it will go. “My hope is that we're creating the next billion-people AI entertainment platform,” Anand says. “Will that look like microdramas or interactive AI games or a hybrid of those? We don’t know yet.”

What else

BitTorrent turns 25. “My plan was not to start a business,” says BitTorrent’s inventor Bram Cohen. “My plan was to start a revolution.” My latest for The Verge.

Comcast’s Sky is acquiring ITV. The deal, if approved, would make Sky Britain’s second-largest broadcaster.

Netflix is betting on premium short-form video. The streamer struck deals with several major publishers, including BuzzFeed Studios, Condé Nast, Hearst Magazines, Penske Media, and People Inc., to add videos as short as three minutes to its catalog.

Midjourney wants studios to reveal their AI use. As part of its defense in a copyright infringement lawsuit, the AI company wants Hollywood to talk about its own use of AI.

More details on recent cuts at Sonos. People laid off include some senior design and product management executives, according to Bloomberg.

Looks like linear TV is popular … on Netflix? Netflix recently began streaming TF1’s linear feed in France. Three weeks later, TF1 said the partnership has already surpassed viewing targets set for the first 18 months.

Meta previews new AI video generation model. Muse video is supposedly better than Veo and Sora.

That’s it

I’ve been a bit distracted lately because of the World Cup, and after watching the Egypt vs. Argentina game, my conclusion is: This has all gone too VAR.

Thanks for reading, have a great weekend!

Cover image courtesy of Shortical.