A24 walked away with trophies in seven categories at the 2023 Academy Awards. "Everything Everywhere All at Once" turned a $14 million budget into $69 million at the global box office, rewriting the rulebook for independent studios on everything from how films get made to how they get marketed. A24 greenlights, at most, around ten projects a year. Which of the thousands of scripts it sees gets picked, and which director gets handed the reins — that consistent aesthetic judgment has been the studio's brand. So when word came that Google DeepMind was investing $75 million in the studio to co-develop AI filmmaking tools, what caught the industry's attention wasn't the price tag. It was the name A24 — a studio widely assumed to be one of the ones keeping AI at arm's length, choosing to partner up instead.

How the creative industry should read this deal hinges on one question: is A24 simply handing over training data, or something else? Look at what DeepMind actually wanted out of the arrangement, and the deal reads very differently.

What DeepMind Wanted Was Editorial Judgment

According to TechCrunch, the two companies are jointly building AI-powered filmmaking tools. DeepMind brings the technical horsepower; A24 brings its production pipeline and aesthetic judgment. The $75 million goes toward building those tools.

Tech companies funding content companies is nothing new. But this deal isn't like Netflix taking a stake in a studio, or Amazon buying MGM. If the goal were securing distribution channels or IP assets, acquiring a production company or licensing content would have done the job. What DeepMind chose instead was co-development — direct access to A24's production floor and its decision-making process.

There's a reason for that. The more technically polished AI film-generation tools get, the harder they run into a bottleneck: aesthetic judgment. Tools like Sora or Runway can spin up a few seconds of video from a text prompt alone. But whether a scene is cinematically alive, which editing rhythm builds an emotional arc, which line of dialogue establishes a character — an AI model has a hard time evaluating any of that on its own. What A24 has built up over two decades in the independent film ecosystem is exactly that judgment: an organizational instinct, sharpened over hundreds of calls, about which director should get which project, which script will hold an audience in a theater seat. Over the runtime of a single film, an editor makes thousands of decisions. Those decision patterns can generate the feedback signal an AI model needs in training. DeepMind isn't investing in movies themselves — it's investing in the standard that determines which movies survive.

The Worry That "A24's Independence Is at Stake" Isn't a Small One

This deal can't be read as purely positive. Look at where A24's competitive edge actually comes from, and you can see how close the DeepMind deal sits to that very edge.

A24 is a studio that chose aesthetic autonomy over funding from a major distributor or streaming platform. Films like "Midsommar" or "Ex Machina" — ones that prioritized authorial vision over commercial appeal — are what put A24 where it is today. What happens when this studio decides "we're making this project," and an outside funder's needs get a seat at that table? Once DeepMind's $75 million is in the mix, the independence of production decisions is at risk of erosion. The question "does this scene work as training data" could start colliding with the question "does this scene serve the film." A funder's needs bending the direction of content production is a pattern that shows up again and again in creative partnerships.

There are copyright and creator-rights concerns here too. One of the central issues in the 2023 Writers Guild of America (WGA) strike was contract language around AI tool use — writers pushed back against their scripts being used to train AI models without consent, and against being forced into arrangements where they simply polish AI-generated output. What terms govern how the A24–DeepMind deal turns writers' and directors' creative judgment into data hasn't been disclosed. Voices in the film criticism community are already warning that this deal could dilute A24's curatorial principles over the long run. That opacity itself is the real basis for the concern.

What Content Directors Should Watch For in a World Where Taste Becomes Data

It's worth spelling out what this deal actually means for solo entrepreneurs and content directors working in Korea.

First, it's worth checking where exactly AI is crossing the line in creative work. Editing assistance, script drafting, image generation — these are already in everyday use. The DeepMind–A24 deal is a signal from the next layer up: AI is moving past mere production assistance and into building the capability to judge what counts as good content in the first place. We already see AI-driven prediction at work in things like forecasting click-through rates on YouTube thumbnails or open rates on newsletter subject lines. The DeepMind–A24 project is an attempt to apply that same logic to film — a far more complex, far longer-form kind of content.

It's also worth watching how ownership of taste data is shifting. A creator's editorial instinct and sense of taste have, until now, been personal, tacit knowledge. Once studios start signing deals that convert that instinct into AI training data, content creators need to start reading the data-use clauses of the AI tools they themselves use — Adobe Firefly, Canva's AI features, video editing tools, all of it. Many generative AI services already include clauses that let them use uploaded files and editing decisions to improve their models. Using your own taste with full knowledge of where it's going is different from using it blind.

Putting your own judgment criteria into words is also practical preparation. If you can explain why you chose this particular scene, or why this particular direction serves the reader, you can judge what's better between an AI-generated draft and your own work. Articulating your standards makes a concrete difference: the more explicit your criteria, the more specific your prompts to AI tools become, and the better the output gets. Just as the way A24's directors describe their own cinematic language can become a meaningful training signal, an individual creator's vocabulary for their own taste is a variable that determines how well they can use AI tools.

It's also worth tracking how fast AI tools are reshaping the cost structure of content production. Filmmaking tools backed by an institution the size of DeepMind will likely be commercially accessible within three to five years. Post-production work, subtitle translation, background music licensing — the costs a Korean solo creator spends producing a single video are exactly the areas these tools are aiming to cut. Whether that's an opportunity or a threat depends on what the creator's actual competitive edge is. For creators competing on cost efficiency, it's a threat. For creators competing on taste and judgment, it's an opportunity — production tools just got cheaper.

There's an interesting paradox here. The more AI tools boost production efficiency, the more differentiation between outputs comes down to knowing which ones are actually good. If you can generate a thousand images an hour, your edge comes from the instinct to pick the right ten. It's a direction that experts studying the future of work have been flagging for a while: the more AI takes over repetitive production, the narrower — and sharper — human judgment and taste become as a competitive edge.

DeepMind didn't choose A24 for its filmmaking technology — it chose it for A24's ability to judge which films are actually alive. Now that studios are starting to hand that judgment over to AI, the competitive ground still left for individual creators is building that judgment up, internally, for themselves.