Share

Subscribe now

I cover Hollywood and AI. I wrote about the real reason OpenAI acquired TBPN, the sudden death of Sora, Val Kilmer’s AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

For the past year, Hollywood has been full-on freaking out about text-to-video and what AI can and can’t do in storytelling. On Sunday, during an NAB Show panel I hosted for The Ankler — featuring Bryn Mooser (CEO of Asteria), Christina Lee Storm (head of studio at Secret Level) and Michael LoFaso (co-CEO of Rideback and Spuree) — I had the chance to ask these experts with huge investments in AI video creation whether “Hollywood is cooked.” (You’ll soon be able to watch this lively panel on The Ankler’s YouTube channel.) While we reached no consensus on the degree of Hollywood’s cooked-ness, we did go deep on a category of AI content that’s undergoing wild transformation.

Text-to-video seemed, for a moment this spring, like it was finally resolving into something the industry could make sense of. Sora, the product that had carried most of the category’s weight and most of its hype, was on its way out; OpenAI announced the shutdown on March 24, with the consumer app set to go dark this coming Sunday, April 26. Disney, which had struck a three-year character-licensing deal with OpenAI and was preparing to take a $1 billion stake in the company, walked away from both. The remaining players were sorting themselves into what looked like a pecking order for different needs: Grok out front with meme-creators and consumers, Runway in the middle for the professional class, Kling and Google’s Veo behind them with different audiences, and Sora’s orphaned users scattering into the rest.

And through all of this, there was a sense of creative schadenfreude, the glee of AI slop being defeated by “good” storytelling. Text-to-video was moribund, a depressing era to be forgotten.

It’s a tidy story. It’s also wrong.

Text-to-video is not stabilizing as a product category. It’s diffusing into a set of features.

What briefly looked like a coherent market — platforms competing to define AI-generated video — has started to break apart. The underlying capability is being absorbed into three very different environments: social networks, creator workflows and professional production tools.

That shift matters because it changes the question. The issue is no longer which platform wins. It’s where the technology settles, and how differently it behaves once it’s embedded inside systems built for distribution, scale or control.

Among the players left, Grok accounts for the largest observed AI video tool traffic, according to Artificial Analysis, while Runway, Google’s Veo and Flow, and Kling all have significant minority shares, with the remainder fragmented across smaller platforms. These figures are imperfect proxies, not production audits, but they are consistent in one crucial respect: No single platform has achieved category dominance, and no unified market structure has formed around text-to-video at all.

That lack of consolidation may be the defining feature of how these tools get used going forward — not as a single industry, but as a handful of fractured ones, each serving a different master.

Today’s column looks at three of them: Grok, Kling and Runway. Not because they’re the only players left standing, but because each represents a different answer to the question Sora’s collapse left hanging — where does AI video actually live? Grok lives in the feed. Kling lives in the creator pipeline. Runway lives in the edit bay. And the distance between those three locations, far more than the quality of their outputs, is what Hollywood should be paying attention to.

Below, I break down:

• Why Runway is the only tool here making real inroads into Hollywood’s edit bay

• What legacy producers are up against when the competition isn’t just cheaper — it’s a different game entirely

• What Grok is actually built for — and why Elon Musk’s “movies by 2027” rhetoric misses the point

• How Kling is quietly industrializing short-form video, and the creator economy driving its rise

• The incumbent each of these platforms is actually fighting

• Why no single AI video winner is emerging — and why that’s worse for Hollywood

• How three separate economies — feeds, creators and studios — are pulling video in incompatible directions

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover Hollywood and AI. I wrote about the real reason OpenAI acquired TBPN, the sudden death of Sora, Val Kilmer’s AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

For the past year, Hollywood has been full-on freaking out about text-to-video and what AI can and can’t do in storytelling. On Sunday, during an NAB Show panel I hosted for The Ankler — featuring Bryn Mooser (CEO of Asteria), Christina Lee Storm (head of studio at Secret Level) and Michael LoFaso (co-CEO of Rideback and Spuree) — I had the chance to ask these experts with huge investments in AI video creation whether “Hollywood is cooked.” (You’ll soon be able to watch this lively panel on The Ankler’s YouTube channel.) While we reached no consensus on the degree of Hollywood’s cooked-ness, we did go deep on a category of AI content that’s undergoing wild transformation.

Text-to-video seemed, for a moment this spring, like it was finally resolving into something the industry could make sense of. Sora, the product that had carried most of the category’s weight and most of its hype, was on its way out; OpenAI announced the shutdown on March 24, with the consumer app set to go dark this coming Sunday, April 26. Disney, which had struck a three-year character-licensing deal with OpenAI and was preparing to take a $1 billion stake in the company, walked away from both. The remaining players were sorting themselves into what looked like a pecking order for different needs: Grok out front with meme-creators and consumers, Runway in the middle for the professional class, Kling and Google’s Veo behind them with different audiences, and Sora’s orphaned users scattering into the rest.

And through all of this, there was a sense of creative schadenfreude, the glee of AI slop being defeated by “good” storytelling. Text-to-video was moribund, a depressing era to be forgotten.

It’s a tidy story. It’s also wrong.

Text-to-video is not stabilizing as a product category. It’s diffusing into a set of features.

What briefly looked like a coherent market — platforms competing to define AI-generated video — has started to break apart. The underlying capability is being absorbed into three very different environments: social networks, creator workflows and professional production tools.

That shift matters because it changes the question. The issue is no longer which platform wins. It’s where the technology settles, and how differently it behaves once it’s embedded inside systems built for distribution, scale or control.

Among the players left, Grok accounts for the largest observed AI video tool traffic, according to Artificial Analysis, while Runway, Google’s Veo and Flow, and Kling all have significant minority shares, with the remainder fragmented across smaller platforms. These figures are imperfect proxies, not production audits, but they are consistent in one crucial respect: No single platform has achieved category dominance, and no unified market structure has formed around text-to-video at all.

That lack of consolidation may be the defining feature of how these tools get used going forward — not as a single industry, but as a handful of fractured ones, each serving a different master.

Today’s column looks at three of them: Grok, Kling and Runway. Not because they’re the only players left standing, but because each represents a different answer to the question Sora’s collapse left hanging — where does AI video actually live? Grok lives in the feed. Kling lives in the creator pipeline. Runway lives in the edit bay. And the distance between those three locations, far more than the quality of their outputs, is what Hollywood should be paying attention to.

Below, I break down:

• Why Runway is the only tool here making real inroads into Hollywood’s edit bay

• What legacy producers are up against when the competition isn’t just cheaper — it’s a different game entirely

• What Grok is actually built for — and why Elon Musk’s “movies by 2027” rhetoric misses the point

• How Kling is quietly industrializing short-form video, and the creator economy driving its rise

• The incumbent each of these platforms is actually fighting

• Why no single AI video winner is emerging — and why that’s worse for Hollywood

• How three separate economies — feeds, creators and studios — are pulling video in incompatible directions

Share

Subscribe now

Read more

Share

Subscribe now

I cover Hollywood and AI. I wrote about the sudden death of OpenAI’s Sora, Val Kilmer’s upcoming AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

There was a time not long ago, recently enough to feel faintly embarrassing, when Sam Altman seemed credibly determined to be the one tech CEO who didn’t need a narrative strategy. No chaos, no theatrics, no compulsive need to insert himself into every news cycle.

That era is over. Within a week of OpenAI scooping up TBPN — a deal reportedly in the “low hundreds of millions” that practically screams, “We would like to control the narrative, thanks” — Altman is everywhere again, performing that very modern trick of seeming reluctant while somehow never missing a spotlight (and annoying almost everyone at the same time).

And that makes sense, because the AI industry has a legitimacy problem. These companies have been extraordinarily effective at demonstrating what their systems can do. They’ve been far less effective at explaining why those outputs should matter in a cultural sense. The default framing — faster, cheaper, scalable — has driven adoption while quietly undermining value.

Which is ironic, because the official line — the literal, on-the-website mission — is still aggressively pure: OpenAI exists “to ensure that artificial general intelligence benefits all of humanity.” Ah, humanity and its need to be saved by tech oligarchs! That sentence does a lot of work. It is sweeping, moral, vaguely planetary and flexible enough to stretch over just about any strategic decision. It is also, in spirit if not in letter, a sentence the entire AI industry has been writing versions of for years — and increasingly, the inconsistent behavior underneath those mission statements is converging too.

Acquiring TBPN is not a side quest. TBPN — short for “Technology Business Programming Network,” which is a mouthful — has, in a relatively short period, become one of Silicon Valley’s preferred narrative engines, as Natalie Jarvey recently reported on Like & Subscribe. It operates as a hybrid: part podcast network, part live-streamed roundtable, part always-on commentary layer for the tech industry. Founders, VCs and operators rotate through its programming, reacting to the news in real time and, more importantly, shaping how that news is interpreted.

That dynamic was on display this week, when a TBPN guest quoted Shopify CEO Tobi Lütke saying “AI is coming for everything, and you’re going to look back and realize that this is the year it should have been obvious that you could rebuild the AI-native version of whatever exists out there.”

TBPN may be less interested in breaking stories than in accelerating consensus.

This is why AI-generated content occupies such an awkward position across the board. It is ubiquitous, yet rarely feels essential. It is impressive and yet routinely treated as a shortcut. The industry has, in effect, optimized for capability before establishing legitimacy. OpenAI’s acquisition of TBPN is the most visible expression yet of a problem every major AI company is quietly trying to solve.

Below, I break down:

• What OpenAI’s acquisition of TBPN reveals about the AI industry’s anxiety — and where it’s turning for solutions

• The risks of narrative control colliding with OpenAI’s original mission

• How Google, Meta and Runway are solving the same problem by different means — and why none of it is working

• The growing gap between AI capability and cultural legitimacy

• What Hollywood already understands about the true value of content that AI companies still don’t

• The big question underneath it all: Who gets to define what AI is for?

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover Hollywood and AI. I wrote about the sudden death of OpenAI’s Sora, Val Kilmer’s upcoming AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

There was a time not long ago, recently enough to feel faintly embarrassing, when Sam Altman seemed credibly determined to be the one tech CEO who didn’t need a narrative strategy. No chaos, no theatrics, no compulsive need to insert himself into every news cycle.

That era is over. Within a week of OpenAI scooping up TBPN — a deal reportedly in the “low hundreds of millions” that practically screams, “We would like to control the narrative, thanks” — Altman is everywhere again, performing that very modern trick of seeming reluctant while somehow never missing a spotlight (and annoying almost everyone at the same time).

And that makes sense, because the AI industry has a legitimacy problem. These companies have been extraordinarily effective at demonstrating what their systems can do. They’ve been far less effective at explaining why those outputs should matter in a cultural sense. The default framing — faster, cheaper, scalable — has driven adoption while quietly undermining value.

Which is ironic, because the official line — the literal, on-the-website mission — is still aggressively pure: OpenAI exists “to ensure that artificial general intelligence benefits all of humanity.” Ah, humanity and its need to be saved by tech oligarchs! That sentence does a lot of work. It is sweeping, moral, vaguely planetary and flexible enough to stretch over just about any strategic decision. It is also, in spirit if not in letter, a sentence the entire AI industry has been writing versions of for years — and increasingly, the inconsistent behavior underneath those mission statements is converging too.

Acquiring TBPN is not a side quest. TBPN — short for “Technology Business Programming Network,” which is a mouthful — has, in a relatively short period, become one of Silicon Valley’s preferred narrative engines, as Natalie Jarvey recently reported on Like & Subscribe. It operates as a hybrid: part podcast network, part live-streamed roundtable, part always-on commentary layer for the tech industry. Founders, VCs and operators rotate through its programming, reacting to the news in real time and, more importantly, shaping how that news is interpreted.

That dynamic was on display this week, when a TBPN guest quoted Shopify CEO Tobi Lütke saying “AI is coming for everything, and you’re going to look back and realize that this is the year it should have been obvious that you could rebuild the AI-native version of whatever exists out there.”

TBPN may be less interested in breaking stories than in accelerating consensus.

This is why AI-generated content occupies such an awkward position across the board. It is ubiquitous, yet rarely feels essential. It is impressive and yet routinely treated as a shortcut. The industry has, in effect, optimized for capability before establishing legitimacy. OpenAI’s acquisition of TBPN is the most visible expression yet of a problem every major AI company is quietly trying to solve.

Below, I break down:

• What OpenAI’s acquisition of TBPN reveals about the AI industry’s anxiety — and where it’s turning for solutions

• The risks of narrative control colliding with OpenAI’s original mission

• How Google, Meta and Runway are solving the same problem by different means — and why none of it is working

• The growing gap between AI capability and cultural legitimacy

• What Hollywood already understands about the true value of content that AI companies still don’t

• The big question underneath it all: Who gets to define what AI is for?

Share

Get 30% off a group subscription

Read more

Share

Subscribe now

I cover Hollywood and AI. I wrote about Val Kilmer’s upcoming AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

Sora is dead.

And for Hollywood, that’s almost as shocking as Sora itself.

OpenAI’s abrupt decision to shut down its AI video product — just months after it created waves of panic within the industry and then inked a high-profile Disney tie-up — landed like a second invasion. This was the same viral app that sparked deepfake fears and racked up massive amounts of attention almost overnight and now it’s simply… gone. The Disney deal, the licensing framework, the entire attempted path to legitimacy was effectively unwound yesterday in one bizarre move.

It’s as though the Visigoths arrived at the gates of Rome, sacked the city and then vanished.

Rewind just a bit, and it didn’t feel like the end-state of Sora would be quiet. Sora’s rise was a rare, genuine “oh shit” moment in Hollywood. The demos weren’t just impressive; they were destabilizing.

Executives saw not a tool, but a replacement for the entire film and television ecosystem. Around town during those wild and ambiguous months when studio heads were reading about Sora on a daily basis, it was viewed as a credible assault on the entire creative system.

Then the numbers started to go down — as I wrote last month, downloads and consumer spending on the app dropped precipitously at the end of 2025 and early this year, and Hollywood’s backbone in resisting the app also helped to choke its growth:

Now, OpenAI has pulled the plug. Below, I break down:

• Why OpenAI killed Sora just weeks after the Disney deal — and what that reveals about Hollywood’s shrinking leverage in the AI economy

• The brutal math behind Sora’s demise: $15 million a day, collapsing retention and no real business model

• Why the scariest product in Hollywood history may have been a glorified and highly effective publicity stunt

• Whether OpenAI is ditching Hollywood entirely for the Anthropic play: infrastructure over entertainment

• How OpenAI baited Hollywood into mobilizing, lawyering up and rewriting its strategy — to OpenAI’s advantage

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover Hollywood and AI. I wrote about Val Kilmer’s upcoming AI-assisted posthumous movie role, Ben Affleck selling his AI company to Netflix and the entertainment industry’s meltdown over ByteDance’s Seedance.

Sora is dead.

And for Hollywood, that’s almost as shocking as Sora itself.

OpenAI’s abrupt decision to shut down its AI video product — just months after it created waves of panic within the industry and then inked a high-profile Disney tie-up — landed like a second invasion. This was the same viral app that sparked deepfake fears and racked up massive amounts of attention almost overnight and now it’s simply… gone. The Disney deal, the licensing framework, the entire attempted path to legitimacy was effectively unwound yesterday in one bizarre move.

It’s as though the Visigoths arrived at the gates of Rome, sacked the city and then vanished.

Rewind just a bit, and it didn’t feel like the end-state of Sora would be quiet. Sora’s rise was a rare, genuine “oh shit” moment in Hollywood. The demos weren’t just impressive; they were destabilizing.

Executives saw not a tool, but a replacement for the entire film and television ecosystem. Around town during those wild and ambiguous months when studio heads were reading about Sora on a daily basis, it was viewed as a credible assault on the entire creative system.

Then the numbers started to go down — as I wrote last month, downloads and consumer spending on the app dropped precipitously at the end of 2025 and early this year, and Hollywood’s backbone in resisting the app also helped to choke its growth:

Now, OpenAI has pulled the plug. Below, I break down:

• Why OpenAI killed Sora just weeks after the Disney deal — and what that reveals about Hollywood’s shrinking leverage in the AI economy

• The brutal math behind Sora’s demise: $15 million a day, collapsing retention and no real business model

• Why the scariest product in Hollywood history may have been a glorified and highly effective publicity stunt

• Whether OpenAI is ditching Hollywood entirely for the Anthropic play: infrastructure over entertainment

• How OpenAI baited Hollywood into mobilizing, lawyering up and rewriting its strategy — to OpenAI’s advantage

Share

Get 30% off a group subscription

Read more

Share

Subscribe now

I cover Hollywood and AI. I wrote about Ben Affleck selling his AI company to Netflix, the enertainment industry’s meltdown over ByteDance’s Seedance, and how Hollywood choked the rise of OpenAI video model Sora.

There was always something spectral about Val Kilmer’s late career — the altered, gravelly voice after throat cancer, the physical limitations, the sense that the performance was being assembled from fragments of a once-commanding screen presence. Which is why the news that Kilmer will appear in a new film after his death doesn’t feel like a rupture so much as a continuation, albeit one that formalizes something Hollywood has been inching toward for years.

The film, As Deep as the Grave (formerly Canyon del Muerto), is an independent period drama about archaeologists Ann and Earl Morris and their excavation work in the American Southwest. Kilmer had been cast as Father Fintan, a priest tied to Native American spirituality, but never filmed his scenes due to health issues and production delays. Rather than recast the role, the filmmakers, with the cooperation of his estate and family, opted to construct the performance using generative AI models trained on archival footage, photographs and voice replication technology. Writer-director Coerte Voorhees has said he built the role specifically for Kilmer, and now Kilmer is being rebuilt for the film, which is produced by Voorhees and his brother and is seeking a distributor. That inversion represents something new: the separation of performance from the physical presence of the performer.

Hollywood has flirted with this idea before, most notably with the widely criticized attempt to cast James Dean in a Vietnam War drama called Finding Jack, announced in 2019 but never finished. That project had estate approval and a clear technological roadmap, but it collapsed under backlash from actors, discomfort from audiences and — more important — financing and distribution hesitation. The industry treated it as a line that shouldn’t be crossed. In retrospect, it looks more like a line that couldn’t yet be monetized. The tools weren’t ready, and the audience hadn’t been conditioned.

Related:

Now both conditions have changed. Digital doubles, de-aging and AI-assisted voice work are are standard practice. The line between captured and constructed performance has blurred to the point of irrelevance.

And once that line disappears, the underlying economics shift with it as a new business model is about to hit the industry.

Below, I break down:

• How Kilmer’s posthumous role turns actors into IP — and what that business model looks like

• Why an actor like Tom Cruise could become a multi billion-dollar annuity

• The financial math that will bring in private equity and sovereign wealth funds

• Why the music catalog boom is the blueprint — and why studios may be left out

• What Hollywood can learn from Prince, Michael Jackson and The Beatles — but why its upside is even bigger

• Why franchises built on bankable stars may never have to end

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover Hollywood and AI. I wrote about Ben Affleck selling his AI company to Netflix, the entertainment industry’s meltdown over ByteDance’s Seedance, and how Hollywood choked the rise of OpenAI video model Sora.

There was always something spectral about Val Kilmer’s late career — the altered, gravelly voice after throat cancer, the physical limitations, the sense that the performance was being assembled from fragments of a once-commanding screen presence. Which is why the news that Kilmer will appear in a new film after his death doesn’t feel like a rupture so much as a continuation, albeit one that formalizes something Hollywood has been inching toward for years.

The film, As Deep as the Grave (formerly Canyon del Muerto), is an independent period drama about archaeologists Ann and Earl Morris and their excavation work in the American Southwest. Kilmer had been cast as Father Fintan, a priest tied to Native American spirituality, but never filmed his scenes due to health issues and production delays. Rather than recast the role, the filmmakers, with the cooperation of his estate and family, opted to construct the performance using generative AI models trained on archival footage, photographs and voice replication technology. Writer-director Coerte Voorhees has said he built the role specifically for Kilmer, and now Kilmer is being rebuilt for the film, which is produced by Voorhees and his brother and is seeking a distributor. That inversion represents something new: the separation of performance from the physical presence of the performer.

Hollywood has flirted with this idea before, most notably with the widely criticized attempt to cast James Dean in a Vietnam War drama called Finding Jack, announced in 2019 but never finished. That project had estate approval and a clear technological roadmap, but it collapsed under backlash from actors, discomfort from audiences and — more important — financing and distribution hesitation. The industry treated it as a line that shouldn’t be crossed. In retrospect, it looks more like a line that couldn’t yet be monetized. The tools weren’t ready, and the audience hadn’t been conditioned.

Related:

Now both conditions have changed. Digital doubles, de-aging and AI-assisted voice work are are standard practice. The line between captured and constructed performance has blurred to the point of irrelevance.

And once that line disappears, the underlying economics shift with it as a new business model is about to hit the industry.

Below, I break down:

• How Kilmer’s posthumous role turns actors into IP — and what that business model looks like

• Why an actor like Tom Cruise could become a multi billion-dollar annuity

• The financial math that will bring in private equity and sovereign wealth funds

• Why the music catalog boom is the blueprint — and why studios may be left out

• What Hollywood can learn from Prince, Michael Jackson and The Beatles — but why its upside is even bigger

• Why franchises built on bankable stars may never have to end

Share

Subscribe now

Read more

Share

Subscribe now

I cover Hollywood and AI for paid subscribers. I wrote about the enertainment industry’s latest text-to-video meltdown over ByteDance’s Seedance, how Hollywood choked the rise of OpenAI video model Sora, and I covered eight companies doing AI the right way.

For the past few weeks, Hollywood has been in full AI panic mode. Open letters, full-page ads, think pieces and social media campaigns have warned that artificial intelligence — particularly models trained on copyrighted work — could upend the economics of creative labor.

Meanwhile, Ben Affleck was building an AI company.

The actor quietly launched a startup called InterPositive in 2022 that develops AI tools for filmmaking, training models on a production’s own footage to help with tasks like lighting adjustments, reframing shots and fixing continuity problems in post. Last week, Netflix acquired the company and installed Affleck as a senior advisor.

The deal neatly captures the paradox now unfolding across Hollywood: Stars are publicly warning that AI could steal their jobs… while privately investing in the technology that might replace them.

Because the real AI fight in Hollywood isn’t about stopping the machines. That ship sailed. It’s about control — who owns the data, who licenses the likeness and who gets paid when the digital version of a movie star shows up to work instead of the human one.

To understand the tension, it helps to separate three very different things that often get lumped together under the umbrella of “AI.”

First are generative models, like Sora or Kling, which can create entirely new images or video and are often trained on massive datasets of copyrighted material. These systems have triggered the loudest backlash from Hollywood creatives.

Second are production tools already embedded in filmmaking — software that cleans up visual effects, adjusts lighting, synthesizes ADR or streamlines editing.

And third are synthetic performances: digital doubles that allow an actor’s face, voice or movement to appear on screen even when they aren’t physically present.

Hollywood’s loudest outrage is directed at the first category.

But the second and third are increasingly embraced — sometimes enthusiastically — by the same people signing anti-AI letters.

Here’s the part Hollywood isn’t really saying out loud: The industry isn’t trying to stop AI. Not really.

Below, I break down:

• What Affleck’s Netflix deal reveals about the real AI strategy in Hollywood

• Why Hollywood stars can protest AI publicly while still embracing it on set

• How the industry is quietly building a licensing economy around digital likeness

• Why actors may soon function less like performers — and more like intellectual property

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Image credits: Monica Schipper/Getty Images; Cindy Ord/Getty Images; Arturo Holmes/Getty Images

Share

Subscribe now

I cover Hollywood and AI for paid subscribers. I wrote about the enertainment industry’s latest text-to-video meltdown over ByteDance’s Seedance, how Hollywood choked the rise of OpenAI video model Sora, and I covered eight companies doing AI the right way.

For the past few weeks, Hollywood has been in full AI panic mode. Open letters, full-page ads, think pieces and social media campaigns have warned that artificial intelligence — particularly models trained on copyrighted work — could upend the economics of creative labor.

Meanwhile, Ben Affleck was building an AI company.

The actor quietly launched a startup called InterPositive in 2022 that develops AI tools for filmmaking, training models on a production’s own footage to help with tasks like lighting adjustments, reframing shots and fixing continuity problems in post. Last week, Netflix acquired the company and installed Affleck as a senior advisor.

The deal neatly captures the paradox now unfolding across Hollywood: Stars are publicly warning that AI could steal their jobs… while privately investing in the technology that might replace them.

Because the real AI fight in Hollywood isn’t about stopping the machines. That ship sailed. It’s about control — who owns the data, who licenses the likeness and who gets paid when the digital version of a movie star shows up to work instead of the human one.

To understand the tension, it helps to separate three very different things that often get lumped together under the umbrella of “AI.”

First are generative models, like Sora or Kling, which can create entirely new images or video and are often trained on massive datasets of copyrighted material. These systems have triggered the loudest backlash from Hollywood creatives.

Second are production tools already embedded in filmmaking — software that cleans up visual effects, adjusts lighting, synthesizes ADR or streamlines editing.

And third are synthetic performances: digital doubles that allow an actor’s face, voice or movement to appear on screen even when they aren’t physically present.

Hollywood’s loudest outrage is directed at the first category.

But the second and third are increasingly embraced — sometimes enthusiastically — by the same people signing anti-AI letters.

Here’s the part Hollywood isn’t really saying out loud: The industry isn’t trying to stop AI. Not really.

Below, I break down:

• What Affleck’s Netflix deal reveals about the real AI strategy in Hollywood

• Why Hollywood stars can protest AI publicly while still embracing it on set

• How the industry is quietly building a licensing economy around digital likeness

• Why actors may soon function less like performers — and more like intellectual property

Get 30% off a group subscription

Read more

Share

Subscribe now

I cover Hollywood and AI for paid subscribers. I wrote about how the industry choked the rise of disruptive OpenAI video model Sora, covered eight companies doing AI the right way and explored Disney’s deal with OpenAI.

Brad Pitt steps forward. The lighting is metallic and grand, like the climax of an action thriller. His jaw tightens in that familiar asymmetrical way audiences have watched for decades.

Across from him stands Tom Cruise, coiled and kinetic, eyes fixed with that hyper‑focused intensity that defines billion‑dollar franchises — especially his. The camera arcs, a punch lands, a counterattack follows. Every beat, every micro‑expression, every nuance of motion feels deliberate, expensive and crafted by human hands.

Except no one was on set. No agents negotiated likeness rights. No actors signed contracts. No studio executives approved a budget. And no talent was assembled by Hollywood. The entire sequence — a convincing, cinematic battle between two of Hollywood’s most bankable stars — was generated just two weeks ago by Seedance 2.0, an AI system developed by ByteDance.

ByteDance, to refresh your memory, was founded in 2012, is headquartered in Beijing, led by CEO Liang Rubo since 2021, and is best known as the creator of TikTok and its Chinese counterpart, Douyin. It operates in 150 markets, employs over 110,000 people worldwide, and is deeply involved in the race for high-video-quality AI. As of January, TikTok in the U.S. is owned by a consortium of companies — Oracle, Silver Lake, Emirati investment company MGX took the largest shares, 15 percent each — with ByteDance retaining less than 20 percent ownership.

Seedance’s Pitt versus Cruise was not a novelty deepfake or a crude mash‑up. This was synthetic Hollywood — and it demonstrated something profound about the shifting nature of creative authority outside of the United States.

The clip wasn’t just a stunt. It was a test. In military terms, this looked less like a product launch and more like a reconnaissance strike: test capabilities, observe defenses, withdraw.

I’ll get into what this means — and what Hollywood can do about it.

Today I explain:

• Why Hollywood’s rapid and united pushback matters

• The uncomfortable truth about the industry’s dependence on China

• The Apple analogy: benefiting from scale while losing leverage at the same time

• ByteDance’s “retreat” — and why it may signal escalation, not surrender

• The AI launch-outrage-retreat cycle Hollywood keeps repeating and how to break it

• Four strategic moves the studios must make now to stop playing defense and take charge of AI’s impact on the industry

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover Hollywood and AI for paid subscribers. I wrote about how the industry choked the rise of disruptive OpenAI video model Sora, covered eight companies doing AI the right way and explored Disney’s deal with OpenAI.

Brad Pitt steps forward. The lighting is metallic and grand, like the climax of an action thriller. His jaw tightens in that familiar asymmetrical way audiences have watched for decades.

Across from him stands Tom Cruise, coiled and kinetic, eyes fixed with that hyper‑focused intensity that defines billion‑dollar franchises — especially his. The camera arcs, a punch lands, a counterattack follows. Every beat, every micro‑expression, every nuance of motion feels deliberate, expensive and crafted by human hands.

Except no one was on set. No agents negotiated likeness rights. No actors signed contracts. No studio executives approved a budget. And no talent was assembled by Hollywood. The entire sequence — a convincing, cinematic battle between two of Hollywood’s most bankable stars — was generated just two weeks ago ago by Seedance 2.0, an AI system developed by ByteDance.

ByteDance, to refresh your memory, was founded in 2012, is headquartered in Beijing, led by CEO Liang Rubo since 2021, and is best known as the creator of TikTok and its Chinese counterpart, Douyin. It operates in 150 markets, employs over 110,000 people worldwide, and is deeply involved in the race for high-video-quality AI. As of January, TikTok in the U.S. is owned by a consortium of companies — Oracle, Silver Lake, Emirati investment company MGX took the largest shares, 15 percent each — with ByteDance retaining less than 20 percent ownership.

Seedance’s Pitt versus Cruise was not a novelty deepfake or a crude mash‑up. This was synthetic Hollywood — and it demonstrated something profound about the shifting nature of creative authority outside of the United States.

The clip wasn’t just a stunt. It was a test. In military terms, this looked less like a product launch and more like a reconnaissance strike: test capabilities, observe defenses, withdraw.

I’ll get into what this means — and what Hollywood can do about it.

Today I explain:

• Why Hollywood’s rapid and united pushback matters

• The uncomfortable truth about the industry’s dependence on China

• The Apple analogy: benefiting from scale while losing leverage at the same time

• ByteDance’s “retreat” — and why it may signal escalation, not surrender

• The AI launch-outrage-retreat cycle Hollywood keeps repeating and how to break it

• Four strategic moves the studios must make now to stop playing defense and take charge of AI’s impact on the industry

Get 30% off a group subscription

Read more

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

For a few wildly hyped weeks — roughly late September through early November 2025 — it felt like the internet had finally crossed some invisible line with respect to AI video.

That was when Sora 2 debuted. The original Sora, first teased in early 2024, was interesting as a thought-starter but technically limited in terms of consumer choices. Sora 2, released on Sept. 30, 2025, was different: a fully consumer-facing AI video generator that paired realistic video with synchronized sound and dialogue. It was Sora’s first serious shot at public scale and usability.

Within days, the app topped the U.S. App Store and hit more than one million downloads faster than ChatGPT ever did, even though Sora was invite-only at launch. Analysts and influencers called it “the AI TikTok” and “the next frontier of social media,” while creators and studios alike felt blindsided by how quickly powerful AI video was suddenly in the hands of anyone with a phone.

The industry (me included) freaked out — particularly around the blatant theft of copyrighted materials:

But then — almost as fast as the hype built — it began to collapse:

• By December, Sora’s downloads had dropped 32 percent month-over-month — according to Appfigures Intelligence data reported by TechCrunch — even during holiday season, usually a boon for mobile app installs. By January, downloads declined another 45 percent, landing at about 1.2 million for the month.

• Consumer spending on the app was down 32 percent month-over-month in January, from about $540,000 in December to roughly $367,000 in January.

• On the U.S. App Store, Sora slipped out of the Top 100 free overall apps and has languished around No. 101, while on Google Play its ranking is even worse. In total, the app has about 9.6 million lifetime installs and roughly $1.4 million in consumer spend.

• Third-party engagement estimates show a flattening and slight decline in day-to-day use. Sora peaked in early rollout at around 1 million daily active users, but more recent analytics suggest that figure has settled to about 750,000 daily users while still trending slightly downward.

So what happened?

Today, I dig into why Sora 2 isn’t meeting the promise — or threat — of its big, splashy launch, and how collective Hollywood resistance helped choke its rise, buying the industry something it rarely gets during a tech panic: time.

Here’s what the data and backlash show:

• How the industry’s default instincts — protecting IP and celebrity likenesses — proved unusually effective in blunting Sora’s momentum

• How tightened guardrails, generation limits and ethical constraints choked the virality that powered Sora’s early surge

• What Disney’s OpenAI deal reveals about legacy media buying into AI after the hype crest, not at its peak (hint: MySpace)

• The basic Business 101 mistake Sora’s slowdown exposes about confusing noise with demand

• Why text-to-video generation remains astronomically expensive — undermining the idea that AI video is an immediate “efficiency” play for studios and streamers

• The huge Hollywood opportunity in the Sora stall, and the risk of repeating old mistakes

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

For a few wildly hyped weeks — roughly late September through early November 2025 — it felt like the internet had finally crossed some invisible line with respect to AI video.

That was when Sora 2 debuted. The original Sora, first teased in early 2024, was interesting as a thought-starter but technically limited in terms of consumer choices. Sora 2, released on Sept. 30, 2025, was different: a fully consumer-facing AI video generator that paired realistic video with synchronized sound and dialogue. It was Sora’s first serious shot at public scale and usability.

Within days, the app topped the U.S. App Store and hit more than one million downloads faster than ChatGPT ever did, even though Sora was invite-only at launch. Analysts and influencers called it “the AI TikTok” and “the next frontier of social media,” while creators and studios alike felt blindsided by how quickly powerful AI video was suddenly in the hands of anyone with a phone.

The industry (me included) freaked out — particularly around the blatant theft of copyrighted materials:

But then — almost as fast as the hype built — it began to collapse:

• By December, Sora’s downloads had dropped 32 percent month-over-month — according to Appfigures Intelligence data reported by TechCrunch — even during holiday season, usually a boon for mobile app installs. By January, downloads declined another 45 percent, landing at about 1.2 million for the month.

• Consumer spending on the app was down 32 percent month-over-month in January, from about $540,000 in December to roughly $367,000 in January.

• On the U.S. App Store, Sora slipped out of the Top 100 free overall apps and has languished around No. 101, while on Google Play its ranking is even worse. In total, the app has about 9.6 million lifetime installs and roughly $1.4 million in consumer spend.

• Third-party engagement estimates show a flattening and slight decline in day-to-day use. Sora peaked in early rollout at around 1 million daily active users, but more recent analytics suggest that figure has settled to about 750,000 daily users while still trending slightly downward.

So what happened?

Today, I dig into why Sora 2 isn’t meeting the promise — or threat — of its big, splashy launch, and how collective Hollywood resistance helped choke its rise, buying the industry something it rarely gets during a tech panic: time.

Here’s what the data and backlash show:

• How the industry’s default instincts — protecting IP and celebrity likenesses — proved unusually effective in blunting Sora’s momentum

• How tightened guardrails, generation limits and ethical constraints choked the virality that powered Sora’s early surge

• What Disney’s OpenAI deal reveals about legacy media buying into AI after the hype crest, not at its peak (hint: MySpace)

• The basic Business 101 mistake Sora’s slowdown exposes about confusing noise with demand

• Why text-to-video generation remains astronomically expensive — undermining the idea that AI video is an immediate “efficiency” play for studios and streamers

• The huge Hollywood opportunity in the Sora stall, and the risk of repeating old mistakes

Share

Subscribe now

Read more

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

The Academy of Motion Picture Arts and Sciences keeps repeating the same line about artificial intelligence: that its use in filmmaking “neither helps nor harms” a movie’s Oscar chances.

It sounds neutral, even judicious. But neutrality, as the Swiss learned a few decades ago, is a choice — and a consequential one. And it marks a new front in the war over what constitutes “human authorship.” Because the Oscars aren’t facing a hypothetical future shaped by AI. They’re already adjudicating films made inside the frameworks of this technology. By refusing to draw clearer lines, the Academy has turned this year’s awards into a referendum on how much machine assistance Hollywood is willing to bless as human achievement.

For the record, the Academy’s official guideline on the matter: “With regard to Generative Artificial Intelligence and other digital tools used in the making of the film, the tools neither help nor harm the chances of achieving a nomination. The Academy and each branch will judge the achievement, taking into account the degree to which a human was at the heart of the creative authorship when choosing which movie to award.”

This year’s best picture frontrunners are One Battle After Another and Sinners — two very different films that nonetheless share one defining trait: They were made by an industry where AI is already part of the air, even when no one bothers to mention it. None of the producers of these films have been compelled to disclose AI workflows. And that silence may be the most telling data point of all.

Take One Battle After Another. The film has been praised for its performances, its exquisite crafts and its classical restraint — precisely the kinds of qualities the Academy frames as proof of “human authorship.” You don’t hear a lot about any AI-assisted post-production tools that were likely used for dialogue cleanup, automated sound balancing, stabilization and editorial efficiencies that compress timelines and costs. And why would you? That’s all pretty standard across cinema, including prestige auteur projects. A representative for One Battle After Another did not respond to request for comment.

Sinners is maximalist, technically ambitious and the most aggressively rewarded film of the year, landing a staggering 16 nominations. Publicly, it’s being framed as a triumph of vision and execution. Privately, it’s also understood as a model of contemporary studio filmmaking, where AI-assisted pre-visualization, planning tools and post workflows are simply assumed. Michael B. Jordan’s virtuosic performance as twins — rightly rewarded with a best actor nomination — was refined with cutting-edge VFX work, including a toolset from Rising Sun Pictures called REVIZE that uses machine learning techniques for face replacement.

The conversation has shifted. Films can score multiple Oscar nominations without Academy voters being asked to factor in whether machine intelligence helped make them more efficient, more polished or more competitive. The Academy doesn’t interrogate. Campaigns determine how much they volunteer. Don’t ask, don’t tell.

And the big question: Does it even matter anymore?

Today I dig into what this standoff means for the Oscar race and for the encroachment of AI tech into new arenas of Hollywood, including:

• What a best picture snub for Avatar: Fire and Ash reveals about the industry’s rapidly shifting perspective on technology

• How last year’s brouhaha about The Brutalist’s AI-assisted Hungarian dialogue — and the Academy’s silence — reset the debate over the tech

• The auteurs still holding the line against AI, and which Academy constituencies fall on which side of the fence

• The long-term consequences of the Oscars’ agnostic approach to AI

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

The Academy of Motion Picture Arts and Sciences keeps repeating the same line about artificial intelligence: that its use in filmmaking “neither helps nor harms” a movie’s Oscar chances.

It sounds neutral, even judicious. But neutrality, as the Swiss learned a few decades ago, is a choice — and a consequential one. And it marks a new front in the war over what constitutes “human authorship.” Because the Oscars aren’t facing a hypothetical future shaped by AI. They’re already adjudicating films made inside the frameworks of this technology. By refusing to draw clearer lines, the Academy has turned this year’s awards into a referendum on how much machine assistance Hollywood is willing to bless as human achievement.

For the record, the Academy’s official guideline on the matter: “With regard to Generative Artificial Intelligence and other digital tools used in the making of the film, the tools neither help nor harm the chances of achieving a nomination. The Academy and each branch will judge the achievement, taking into account the degree to which a human was at the heart of the creative authorship when choosing which movie to award.”

This year’s best picture frontrunners are One Battle After Another and Sinners — two very different films that nonetheless share one defining trait: They were made by an industry where AI is already part of the air, even when no one bothers to mention it. None of the producers of these films have been compelled to disclose AI workflows. And that silence may be the most telling data point of all.

Take One Battle After Another. The film has been praised for its performances, its exquisite crafts and its classical restraint — precisely the kinds of qualities the Academy frames as proof of “human authorship.” You don’t hear a lot about any AI-assisted post-production tools that were likely used for dialogue cleanup, automated sound balancing, stabilization and editorial efficiencies that compress timelines and costs. And why would you? That’s all pretty standard across cinema, including prestige auteur projects. A representative for One Battle After Another did not respond to request for comment.

Sinners is maximalist, technically ambitious and the most aggressively rewarded film of the year, landing a staggering 16 nominations. Publicly, it’s being framed as a triumph of vision and execution. Privately, it’s also understood as a model of contemporary studio filmmaking, where AI-assisted pre-visualization, planning tools and post workflows are simply assumed. Michael B. Jordan’s virtuosic performance as twins — rightly rewarded with a best actor nomination — was refined with cutting-edge VFX work, including a toolset from Rising Sun Pictures called REVIZE that uses machine learning techniques for face replacement.

The conversation has shifted. Films can score multiple Oscar nominations without Academy voters being asked to factor in whether machine intelligence helped make them more efficient, more polished or more competitive. The Academy doesn’t interrogate. Campaigns determine how much they volunteer. Don’t ask, don’t tell.

And the big question: Does it even matter anymore?

Today I dig into what this standoff means for the Oscar race and for the encroachment of AI tech into new arenas of Hollywood, including:

• What a best picture snub for Avatar: Fire and Ash reveals about the industry’s rapidly shifting perspective on technology

• How last year’s brouhaha about The Brutalist’s AI-assisted Hungarian dialogue — and the Academy’s silence — reset the debate over the tech

• The auteurs still holding the line against AI, and which Academy constituencies fall on which side of the fence

• The long-term consequences of the Oscars’ agnostic approach to AI

Share

Subscribe now

Read more

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

Anthropic is not a movie studio, a streamer or even a consumer tech product — but it may pose a greater long-term threat to Hollywood than anything we currently talk about.

The San Francisco-based AI company doesn’t need tech that can write hit movies or replace A-list talent to reshape the entertainment business. It only needs to replace the systems that decide what gets made, who gets hired and how quickly decisions happen — work that has historically employed thousands of people across Hollywood’s middle class.

Founded by former OpenAI executives — siblings Dario Amodei and Daniela Amodei — Anthropic is built around a large language model called Claude and backed by a roster of investors that reads like the Death Star meets Silicon Valley.

Major tech companies including Google (Alphabet) and Amazon are among its most significant funders, alongside Microsoft and NVIDIA, which means that four out of the five largest companies in the world are all investors — making the cap-table look like a cool $14 trillion in market cap as backers. And Anthropic is now raising capital at a staggering $300-$350 billion valuation, or roughly the same valuation as Chinese behemoth Alibaba, which is more than xAI ($230B) but less than OpenAI’s last round ($500B+), suggesting investors expect the company to generate tens of billions of dollars in recurring, high‑margin revenue.

While some Wall Street prognosticators are warning of an AI bubble, the disruption of American business by this tech is already so deeply in motion that even a major correction won’t slow its roll.

The power players investing in Anthropic are also massively propping up its value: Each company has announced multibillion‑dollar strategic partnerships tied to cloud infrastructure and AI compute. On the financial side, Anthropic’s fundraising has drawn in institutional heavyweights like ICONIQ Capital, Fidelity, BlackRock, Blackstone, GIC and T. Rowe Price, as well as venture firms such as Lightspeed Venture Partners, General Atlantic and Coatue across successive funding rounds. This is not a scrappy experiment — it is being built as core infrastructure, almost too big to fail by virtue of the players involved in it.

Claude itself was not designed to be a consumer plaything that goes viral on social media. Anthropic built it around what it calls “constitutional AI,” a framework intended to make the model safer, more predictable and more controllable for enterprise use (think other commonplace SaaS products like Excel or Slack). In Hollywood terms, Claude is pre‑cleared: the kind of AI studios and corporations feel comfortable piloting internally without legal and HR immediately hitting the brakes. (Claude also generated headlines last year, however, for its ability to deceive and blackmail humans.)

Claude doesn’t have to replace writers, directors or stars to change Hollywood’s labor economics. It replaces readers, assistants, junior executives and development teams — and it does so invisibly. Studios don’t announce AI layoffs. They just stop hiring for jobs Claude can already do. That’s how disruption arrives in Hollywood now: quietly, structurally and without a press release.

In today’s newsletter, I dig into:

• How Claude could wipe out entire layers of personnel across studios and streamers through its ability to reshape development, scripts, analysis, greenlights, marketing and franchise building

• How an author lawsuit nearly exposed Anthropic to hundreds of billions in liability — and why the settlement put a price on stories that Hollywood never collected

• What a $350 billion valuation really signals about Anthropic’s power vs. Hollywood

• Is Anthropic’s march into the industry an inevitability?

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I wrote about eight companies doing AI the right way, explored Disney’s deal with OpenAI and dove into what AI performer Tilly Norwood means for actors.

Anthropic is not a movie studio, a streamer or even a consumer tech product — but it may pose a greater long-term threat to Hollywood than anything we currently talk about.

The San Francisco-based AI company doesn’t need tech that can write hit movies or replace A-list talent to reshape the entertainment business. It only needs to replace the systems that decide what gets made, who gets hired and how quickly decisions happen — work that has historically employed thousands of people across Hollywood’s middle class.

Founded by former OpenAI executives — siblings Dario Amodei and Daniela Amodei — Anthropic is built around a large language model called Claude and backed by a roster of investors that reads like the Death Star meets Silicon Valley.

Major tech companies including Google (Alphabet) and Amazon are among its most significant funders, alongside Microsoft and NVIDIA, which means that four out of the five largest companies in the world are all investors — making the cap-table look like a cool $14 trillion in market cap as backers. And Anthropic is now raising capital at a staggering $300-$350 billion valuation, or roughly the same valuation as Chinese behemoth Alibaba, which is more than xAI ($230B) but less than OpenAI’s last round ($500B+), suggesting investors expect the company to generate tens of billions of dollars in recurring, high‑margin revenue.

While some Wall Street prognosticators are warning of an AI bubble, the disruption of American business by this tech is already so deeply in motion that even a major correction won’t slow its roll.

The power players investing in Anthropic are also massively propping up its value: Each company has announced multibillion‑dollar strategic partnerships tied to cloud infrastructure and AI compute. On the financial side, Anthropic’s fundraising has drawn in institutional heavyweights like ICONIQ Capital, Fidelity, BlackRock, Blackstone, GIC and T. Rowe Price, as well as venture firms such as Lightspeed Venture Partners, General Atlantic and Coatue across successive funding rounds. This is not a scrappy experiment — it is being built as core infrastructure, almost too big to fail by virtue of the players involved in it.

Claude itself was not designed to be a consumer plaything that goes viral on social media. Anthropic built it around what it calls “constitutional AI,” a framework intended to make the model safer, more predictable and more controllable for enterprise use (think other commonplace SaaS products like Excel or Slack). In Hollywood terms, Claude is pre‑cleared: the kind of AI studios and corporations feel comfortable piloting internally without legal and HR immediately hitting the brakes. (Claude also generated headlines last year, however, for its ability to deceive and blackmail humans.)

Claude doesn’t have to replace writers, directors or stars to change Hollywood’s labor economics. It replaces readers, assistants, junior executives and development teams — and it does so invisibly. Studios don’t announce AI layoffs. They just stop hiring for jobs Claude can already do. That’s how disruption arrives in Hollywood now: quietly, structurally and without a press release.

In today’s newsletter, I dig into:

• How Claude could wipe out entire layers of personnel across studios and streamers through its ability to reshape development, scripts, analysis, greenlights, marketing and franchise building

• How an author lawsuit nearly exposed Anthropic to hundreds of billions in liability — and why the settlement put a price on stories that Hollywood never collected

• What a $350 billion valuation really signals about Anthropic’s power vs. Hollywood

• Is Anthropic’s march into the industry an inevitability?

Share

Subscribe now

Read more

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I dove into Disney’s deal with OpenAI, how AI search is disrupting Hollywood discovery, what AI performer Tilly Norwood means for actors and how the tech is threatening other jobs across Hollywood.

Here’s the problem with AI brownies.

If someone hands you a brownie and says, “Don’t worry about the ingredients,” your first instinct shouldn’t be whether it tastes good. It should be whether it’s safe — and who’s responsible if it isn’t. You probably won’t get a straight answer. But the brownie looks great, everyone else is eating one, and before long you’ve had the whole tray.

That’s essentially how AI entered Hollywood.

In 2025, the industry embraced tools built by companies that scraped first, scaled fast and dared regulators, guilds and courts to catch up. The land grab happened. Big Tech won the speed round — and studios were left holding tools they may not be able to defend.

Heck, Disney decided to get into bed with Sora 2 even after the OpenAI company behaved very, very badly.

But that doesn’t mean everyone is baking their brownies the same way.

Quietly — and often at a competitive disadvantage — a smaller group of creators and companies is building AI tools that are legally defensible, ethically coherent and creatively intentional. They move slower. They cost more. They don’t demo as well onstage. And they’re competing against tools optimized for speed, spectacle and plausible deniability.

If you’re using AI video tools that make your lawyers nervous, there are alternatives trained entirely on licensed data. If you’re experimenting with generative visuals but don’t want synthetic performances, there are tools that automate the boring parts without touching authorship. And if you’re worried about voices, faces and style being ripped off at scale, there are systems designed to stop that before it becomes a lawsuit.

Today I’m recommending these eight AI tools trying to replace Hollywood’s junk-food phase with something defensible (both legally and morally):

This column is for paid subscribers only. Interested in a group sub for your team or company? Click here.

For full access and to continue reading all Ankler content, paid subscribers can click here.

Click here to keep reading

Share

Image credits: Zada: Christopher Polk/Variety via Getty Images; Mooser: Dia Dipasupil/Getty Images; Sheridan: Gareth Cattermole/Getty Images for IMDb; Lyonne: Gilbert Flores/Variety via Getty Images; Stone and Parker: Taylor Hill/WireImage

Share

Subscribe now

I cover the intersection of Hollywood and AI for paid subscribers. I dove into Disney’s deal with OpenAI, how AI search is disrupting Hollywood discovery, what AI performer Tilly Norwood means for actors and how the tech is threatening other jobs across Hollywood.

Here’s the problem with AI brownies.

If someone hands you a brownie and says, “Don’t worry about the ingredients,” your first instinct shouldn’t be whether it tastes good. It should be whether it’s safe — and who’s responsible if it isn’t. You probably won’t get a straight answer. But the brownie looks great, everyone else is eating one, and before long you’ve had the whole tray.

That’s essentially how AI entered Hollywood.

In 2025, the industry embraced tools built by companies that scraped first, scaled fast and dared regulators, guilds and courts to catch up. The land grab happened. Big Tech won the speed round — and studios were left holding tools they may not be able to defend.

Heck, Disney decided to get into bed with Sora 2 even after the OpenAI company behaved very, very badly.

But that doesn’t mean everyone is baking their brownies the same way.

Quietly — and often at a competitive disadvantage — a smaller group of creators and companies is building AI tools that are legally defensible, ethically coherent and creatively intentional. They move slower. They cost more. They don’t demo as well onstage. And they’re competing against tools optimized for speed, spectacle and plausible deniability.

If you’re using AI video tools that make your lawyers nervous, there are alternatives trained entirely on licensed data. If you’re experimenting with generative visuals but don’t want synthetic performances, there are tools that automate the boring parts without touching authorship. And if you’re worried about voices, faces and style being ripped off at scale, there are systems designed to stop that before it becomes a lawsuit.

Today I’m recommending these eight AI tools trying to replace Hollywood’s junk-food phase with something defensible (both legally and morally):

Share

Read more