A24 is the independent studio largely responsible for making modern arthouse movies mainstream, with recent hits like Marty Supreme and Backrooms shaping cultural conversations. 

Now the cult film production company is partnering with Google’s AI-focused DeepMind team to develop AI tools for its filmmakers, according to an exclusive report from the Wall Street Journal. As part of the deal, Google is investing $75 million into A24’s 20-person Labs team to create new tools for movie production and distribution. These tools will be available to A24 creators, but they’ll also funnel back into Google’s AI ecosystem.

Google previously partnered with Darren Aronofsky, a firm believer in using generative AI in filmmaking, and the director of movies like Black Swan and The Whale. But this is the first time the company has staked money in a Hollywood studio.

The search giant currently owns the massive video hosting website, YouTube, and one of the most widely available AI video generation tools, Veo. DeepMind’s investment in A24 represents an expansion of its research scope, which has recently expanded into video games as well.

The deal between the studio and the tech giant doesn’t allow Google to train AI models on A24’s catalog of television shows and movies. Instead, A24 Labs is currently developing an application to produce AI-generated storyboards, the rough draft art created before a movie’s visuals are fully fleshed out.

When asked for comment, a representative of the DeepMind team pointed CNET toward Google’s official blog post about the partnership.

According to that post, Google’s “specific goals, technical outputs and creative milestones of this initiative will evolve over time,” indicating that the scope of this partnership may change as A24 and DeepMind foster a “deep research and development collaboration… spanning multiple projects over time.”

A representative for A24 did not immediately respond to a request for comment.

AI tools in cinema 

The A24 Labs team responsible for creating the new AI tools is headed by Scott Belsky, who left Adobe’s executive team for his new role in January 2025.

According to the Wall Street Journal, Belsky believes that audience and director pushback against AI tools hinges on public perception that generative AI is used to make movies cheaper and faster.

“We think there are better uses that preserve creative control and support risk-taking,” Belsky told the Wall Street Journal. He argued that A24’s AI usage “won’t look anything like the prompted generation type of AI that people feel uncomfortable with.”

While it’s true that actors and directors have by and large rallied against introducing generative AI into the filmmaking process, Hollywood studios have a far more hot-and-cold relationship with the burgeoning industry. Some producers have threatened legal action against AI giants for alleged copyright violations, while others, such as Disney, have attempted to cut ill-fated deals with companies like OpenAI.

A24’s unconventional storytelling 

A24 was founded in 2012 by three film production veterans with a mission to release movies other companies wouldn’t touch. Some of the company’s earliest productions, including Oscar Award-winning pictures Moonlight and Lady Bird, have become cult classics, championed for their portrayal of coming-of-age stories and for featuring characters from marginalized communities that are typically underrepresented in Hollywood films.

While the production company has long been renowned among cinephiles, it arguably became a household name in 2022 with the release of Everything Everywhere All at Once, its first film to surpass $100 million in box-office earnings.

A24’s more recent films have consistently scored critical and financial accolades. Backrooms became the largest original horror movie debut in history, and has raked in more than $270 million at the box office during its theatrical release. A24’s upcoming Elden Ring movie, directed by Alex Garland and adapted from the popular FromSoftware video game, is slated to be the company’s biggest production yet, commanding a $175 million budget.

The DeepMind deal is representative of the cultural clash at the heart of A24’s business model: While the production company elevates independent talent, first-time directors and scripts that other studios might not be willing to take a chance on, it’s still a business worth up to $3.5 billion that’s beholden to investors.

It’s unclear how the Google AI deal may yet affect A24’s sterling public reputation, but it’s very unlikely that the audience of theatergoers that shows up time and time again to support independent projects will be happy with the news.

Spoiler: It’s more Siri stuff.

from the but-think-of-the-children! dept

The Sixth Circuit just handed Ohio a win on its social media law restricting minors, and reading the majority opinion, it’s immediately obvious why: the court fell hook, line, and sinker for the moral panic that social media is inherently poisonous to children. The first few pages of the decision are nothing but cherry-picked, out-of-context horror stories about what social media does to kids — with no mention of the substantial corresponding evidence that the picture is far more complicated than that.

This was, frankly, predictable. Last year, when the Supreme Court heard the FSC v. Paxton case regarding online age verification for adult content websites, we were disappointed that most of the internet industry sat the case out, deciding that they didn’t want to go on the record as supporting the First Amendment for porn sites. We’ve discussed how deciding to sit the case out was a mistake that would come back to bite the companies. When the Supreme Court decided, effectively overturning earlier Supreme Court cases, that age verification laws were fine for porn sites, we also warned that this seemed unlikely to be limited just to adult content. Turns out, the age verification/parental consent virus is spreading.

A bunch of states have passed laws requiring age verification or limiting teenagers’ access to all sorts of non-porn websites, and at least some courts are willing to go along with it. In Ohio, the state passed a very bad social media law a few years ago and NetChoice sued. The district court realized how problematic the law was, including the requirements regarding “parental consent” for kids to use social media, and put the law on hold, which the Sixth Circuit has now reversed.

This new ruling starts out by saying that NetChoice isn’t the proper party to bring this challenge, suggesting that if there is any violation of anyone’s rights, it would be the children who use these social media platforms, and NetChoice represents the companies, not their users. There are some problems with this finding that seem to go against precedent, but if this was all that the court did, a new challenge could be brought on behalf of users.

Instead, though, the majority opinion decides to move forward anyway and consider the First Amendment issue. It properly says that the law must survive strict scrutiny (Ohio tried to argue for lower standards) but then — in a move that can only be described as delusional — claims that the act survives strict scrutiny. If anyone has read the Supreme Court’s ruling in Brown v. EMA — in which the Supreme Court said that California’s law requiring parental consent in the sale of video games was unconstitutional — it should be easy to see how this law is unconstitutional as well. For some reason the Sixth Circuit thinks that this is different, mainly because the majority, again, has misread some of the moral panic studies on social media and kids:

the state’s supporting evidence in Brown was weaker than Yost’s evidence here. In Brown, California “acknowledge[d] that it [could ]not show a direct causal link between violent video games and harm to minors[,]” relied on studies that had “been rejected by every court to consider them” and most of which “suffer[ed] from significant, admitted flaws in methodology[,]” and “show[ed] at best some correlation [with] minuscule real-world effects, such as children’s feeling more aggressive or making louder noises in the few minutes after playing a violent game . . . .” Id. at 799–800. Yost relies on findings by the Surgeon General and researchers who have identified a potentially significant connection between social media and serious harms such as depression and body dysmorphia. Additionally, the evidence in Brown suggested that the effects of violent video games were materially indistinguishable from the effects from other sources of violent material. Id. at 800–01. NetChoice has not brought to our attention evidence that other forms of media have the same negative effects on minors as social media. The link between social media and minors’ wellbeing seems “neither novel nor implausible[,]” and therefore Ohio has met the “quantum of empirical evidence needed to satisfy heightened judicial scrutiny of [its] legislative judgment[] . . . .” OPAWL – Bldg. AAPI Feminist Leadership v. Yost, 118 F.4th 770, 780 (6th Cir. 2024).

So, California lost in Brown (in part) because the state couldn’t show causal evidence of a connection between violent video games and harm, and the Sixth Circuit says social media is different… pointing to the Surgeon General’s report… even though the Surgeon General’s report admits that there is no causal evidence of harm, and says that many kids get tremendous benefit from social media. In particular, that report said that LGBTQ+ kids, especially those in homes with parents who disapprove of their kids’ sexuality, find social media to be a lifeline.

It’s difficult to see how the Sixth Circuit’s majority ruling here makes any sense, unless the majority simply thinks it’s okay for LGBTQ+ and other marginalized kids to suffer. And, really, so much of this ruling makes sense if you read it that way.

The majority opinion, incredibly, brushes off the idea (raised by NetChoice) that it’s not easy for platforms to verify a parent-child relationship. This is a huge problem with any of these laws. How exactly can platforms confirm such a relationship without deeply invasive, privacy-destroying data collection? And, even then, how is a platform to know whether or not a child and parent are estranged, or if there’s a custody battle, or anything of that nature? The majority just dismisses this out of hand, saying it’s easy to figure it out.

First, verifying parental consent will not necessarily present such a great challenge.

That’s literally all it says on that. I’m dying to know exactly how Judge Eric Clay thinks this is possible. Because if he’s figured it out, he’d be the first to do so and he could sell that to every internet platform out there now required to figure out this technical challenge.

There is a dissent from Judge Kevin Ritz, who says that he would affirm the lower court’s ruling, though he would limit the injunction to just NetChoice’s members, rather than against the law as a whole. He notes that NetChoice should have standing to sue, that the law requires strict scrutiny, but that there is no way it should survive strict scrutiny. Unlike his colleagues, who tap dance around Supreme Court precedent by screaming “but think of the children!” Judge Ritz points out that rights are rights. And kids have them.

Yost cannot meet this high bar. At the outset, it is worth emphasizing that “[m]inors are entitled to a significant measure of First Amendment protection, and only in relatively narrow and well-defined circumstances may government bar public dissemination of protected materials to them.” Brown, 564 U.S. at 794 (citation modified). After all, “[a] fundamental principle of the First Amendment is that all persons have access to places where they can speak and listen, and then, after reflection, speak and listen once more.” Packingham v. North Carolina, 582 U.S. 98, 104 (2017). Although the websites the Act seeks to regulate may pose serious risks to children, these sites are also zones for constitutionally protected speech. And they cultivate “vast democratic forums” with the “potential to alter how we think, express ourselves, and define who we want to be.” Id. at 104-05 (citation omitted). Indeed, the Supreme Court has described social media platforms as “perhaps the most powerful mechanisms available to a private citizen to make his or her voice heard.”…

The parental-consent provision at the core of the Act represents a significant burden on the rights of NetChoice’s minor users to avail themselves of these “powerful mechanisms” of speech. Id. For many minor users, in fact, the Act would represent an “insurmountable [ ] barrier to entry for online speech.” Carr, 789 F. Supp. 3d at 1223; see also Uthmeier, 2025 WL 3458571, at *17 (Rosenbaum, J., dissenting) (“When it comes to speaking online, the Act effectively prohibits many minors from speaking at all.”). Like my colleagues and Yost, I do not doubt “that unfettered social media access can and does harm minors.” Griffin, 2025 WL 978607, at 11. But the state does not have “a free-floating power to restrict the ideas to which children may be exposed.” Brown, 564 U.S. at 794. “We should be wary of governments supplanting parents in deciding which ideas children should and should not be exposed to.” Uthmeier, 2025 WL 3458571, at 18 (Rosenbaum, J, dissenting). Besides, the question under strict scrutiny “is not whether [the Act] would be effective, but whether there are less restrictive alternatives.”

And, unlike his colleagues, Ritz notes that the Supreme Court’s ruling in Brown v. EMA appears directly on point:

Brown is instructive. There, the Supreme Court considered a law that prohibited selling or renting violent video games to minors, though minors could still access these games if purchased by an adult. Brown, 564 U.S. at 789, 802. In enacting this law, the legislature created a “wholly new category of content-based regulation that [wa]s permissible only for speech directed at children.” Id. at 794. The Supreme Court “emphatically” rejected this law as “unprecedented and mistaken.” Id. at 792, 794. And although the Court recognized that the legislature’s goal of “addressing a serious social problem” was “legitimate,” the solution was both “seriously underinclusive” and “seriously overinclusive.” Id. at 805.

Ohio has made the same error.

The dissent also calls out the silliness of Clay’s bold assertion that it’s easy for platforms to verify a parent-child relationship noting that, not only is this not true, it’s yet another reason why the law is clearly unconstitutional:

The Court in Brown noted that “one parent (or even an aunt or uncle)” could conceivably purchase a violent video game for a child, but the statute at issue did not have “any requirement[] as to how this parental or avuncular relationship is to be verified; apparently the child’s or putative parent’s, aunt’s, or uncle’s say-so suffices.” 564 U.S. at 802. Here, the Act allows covered operators to register parental consent by one of five ways, including by signing a form, using a credit card, or calling “a toll-free telephone number . . . staffed by trained personnel.” Ohio Rev. Code § 1349.09(B)(1). But Yost cannot explain how these methods prove a parental relationship. For instance, how would an operator know that a person who called the toll-free telephone number is actually the prospective minor user’s parent? Apparently, the caller’s “say-so” is sufficient. Brown, 564 U.S. at 802. The Act’s approach to protecting children fails strict scrutiny.

The dissent also mocks the majority claim that this law somehow protects children. Since the entire point of the law is a one-time consent, Judge Ritz notes that, if we assume that there is real harm from social media to kids, then the second a parent gives consent, that harm is still likely to occur. Ritz calls out the underlying reality that Judge Clay misses: Clay assumes that all parents will refuse to give consent.

In finding that the Act survives strict scrutiny, Judge Clay argues that “Ohio seeks to involve parents because their involvement itself will inherently mitigate some of the harms of unsupervised social media use.” Clay Op. at 23. But it is hard to see how one-time parental consent mitigates such harms unless those parents categorically prohibit their children from using social media on the front end. After all, the Act does not require parents to supervise their children as they use the covered websites. Nor does it create mechanisms for parents to check in with their children about social media habits or to re-consent periodically. The Act only requires that parents are engaged in the beginning and in a binary way—they can either allow their children to use covered websites or not. This approach does not narrowly or effectively advance the goal of involving parents in children’s social media use.

With similar laws popping up all over the country, sooner or later this is absolutely going to the Supreme Court.

When it gets there, here’s what the justices will be looking at: a majority opinion that treats the Surgeon General’s report as proof of causation even though the Surgeon General’s own report explicitly admits there’s no causal evidence of harm. An opinion that waves Brown v. EMA aside while the dissent methodically demonstrates it’s “directly on point.” And an opinion that resolves the question of how platforms are actually supposed to verify parent-child relationships — a problem no one in the industry has managed to solve — with a single sentence asserting it won’t “necessarily present such a great challenge.” No mechanism. No explanation. Just confidence.

Judge Ritz’s dissent does what the majority refuses to: it applies the law. Minors have First Amendment rights. Social media platforms are, as the Supreme Court said in Packingham, “perhaps the most powerful mechanisms available to a private citizen to make his or her voice heard.” And one-time parental consent — which requires no ongoing supervision, verifies nothing meaningful, and defaults to “apparently the caller’s say-so is sufficient” — cannot survive strict scrutiny.

The majority chose moral panic over legal analysis. Now it’s the Supreme Court’s problem.

Filed Under: 6th circuit, eric clay, free speech, kevin ritz, ned yost, ohio, parental consent, social media

Companies: netchoice

Every small business is looking for an edge.

For some, that means protecting margins. For others, it means finding new customers, sourcing better products, entering new markets or simply running the business with less friction.