South Korean science-fiction film Steelay opened in cinemas on 21 October after using AI tools throughout production, The Korea Times reports. The 90-minute feature is a notable test of whether AI-assisted filmmaking can carry a commercial-length story, rather than just supply a striking clip or two.
Watch Desk analysis
What happened
Produced by Laon Company Plus on a budget of around 1 billion won (about US$750,000) over six months, Steelay follows a young man who joins an AI ally to fight a virus and a rogue autonomous robot. The Korea Times describes it as the longest Korean AI feature film to date.
The production team generated roughly 13,000 options and selected the final scenes through human decisions. Human voice actors also dubbed the film, with lip-syncing, sound mixing and other finishing work used to refine the result. The paper notes visible AI-generated elements and occasional awkward action, so this is a reported production milestone, not a claim that the technology has made conventional filmmaking obsolete.
Why it matters
The useful detail is the division of labour: AI helped generate material across the production, while people chose what made the film and handled its performance and finishing. That gives audiences a more concrete case to judge than another grand promise about AI remaking cinema.
A modest budget and six-month production schedule make the project worth watching, but they do not prove that AI alone made it faster or cheaper. The report offers no like-for-like comparison with a conventionally produced film. Still, a feature reaching cinemas is a tougher test than a short online experiment. The awkward bits matter too: commercial storytelling has nowhere to hide when the novelty wears off.
Our read
Steelay is best treated as a real-world production experiment with human editors firmly in the loop. The 13,000 discarded options are less a victory lap for automation than a reminder that generating material and deciding what is worth keeping are different jobs.
What to watch
- How audiences respond to the film beyond the novelty of its AI-assisted production.
- Whether later AI-made features show improvements in action, performance and visual consistency.
- What the production team says about the tools and workflow behind the film.
- Whether similar projects share enough production detail to make meaningful cost or time comparisons.
Discussion spark: If an AI-assisted film needs people to choose from thousands of generated scenes and finish the performances, is that a new filmmaking method or mainly a new set of tools for human editors?
Sources and evidence
- Source update (29 October 2026, 11:39 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.