Alex Proyas takes over the AI-enabled Antonio Banderas epic Cortés after filming began

Cinematographers at Camerimage push back on greybox shoots, where AI adds the lighting after the shoot

At Camerimage, cinematographers say greybox shoots, where actors are filmed on plain sets and AI adds the sets and lighting later, move the look of a film into post. The Bitcoin detail rests on one IndieWire line.

Asked whether he enjoys greybox work, cinematographer Bartosz Nalazek didn't hesitate: "Totally not. It's horrible, I'm brain-dead." For people whose craft is light, the fear at Camerimage this week is that the job shrinks to setting up a blank frame for someone else to finish.

Greybox is the word the festival keeps returning to. Actors are filmed on a plain, neutral set, and AI adds the production design and the lighting afterward. IndieWire reports it was used on Doug Liman's upcoming Bitcoin and is spreading to other projects, and Nalazek says a major TV show he just wrapped shot 40 percent of its scenes that way. Cinematographers describe directors asking for the softest, most generic light on set, not because it looks best but because it leaves the cleanest canvas for changes later.

Nobody in Toruń is pretending the method is useless. It takes weather and locations out of the schedule, and speakers who have worked on these sets point out that the productions still hire hundreds of crew. The festival has drawn its own line, too: Marek Żydowicz, its founder, told The Hollywood Reporter that Camerimage does not accept films generated entirely by AI and favors hybrid films where the live actors decide how far AI may change their images.

Our take: greybox doesn't remove lighting, it moves it to post. Whoever sets the look there owns the image, a seat open to cinematographers and AI creators alike. What nobody at the festival has answered yet is whether the people who light for a living will be the ones sitting in it.

Source: IndieWire, Nov 9, 2026; The Hollywood Reporter and Deadline, Nov 8, 2026. Camerimage runs in Toruń through Nov 14.

Disney names Adobe its AI partner, with private models trained on Disney characters

Adobe will build Firefly Foundry models on Disney characters, environments and styles that only Disney staff can use. No value, term or film title was disclosed, and the parks sketch tool dates from June.

Disney has decided how AI gets near its characters: inside models it owns, and nowhere else. For independent creators that means studio-grade tools behind a locked door, while Disney keeps suing the open tools that trained on its films.

On Monday Disney named Adobe its Official AI Innovation Partner in a multi-year deal. Adobe will build private Firefly Foundry models trained on Disney characters, environments and visual styles, and Reuters reported they will be proprietary to Disney. Neither company disclosed a value, a term or a film title.

The sketch-to-concept-art tool getting the headlines is five months old: Imagineering announced it in June. What is new is the partner title, work with StudioLAB, the Walt Disney Studios innovation hub whose first projects are custom models for post-production and VFX, and projects with the ESPN archive.

Disney's quote came from its head of Corporate Alliances, the team that runs brand partnerships, and Adobe will grow its presence across Disney's media platforms. It reads as much like a sponsorship as a technology choice, though nobody has said which way the money moves.

It also runs the opposite way to Disney's last AI deal. The Sora license pushed characters out to fans until OpenAI shut Sora in March. This one pulls AI inside, where only Disney staff can touch it.

Our take: this is a template for control, not access. Fan work with Disney characters is no safer than it was last week, so build your own visual library and style, because nobody is licensing theirs to you.

Source: Adobe newsroom and Reuters, Nov 9, 2026.

Read the full article: thegenerated.com/stories/disney-adobe-ai-partner-private-models-20261110

Ellison calls AI a creative unlock while Skydance's filing counts it only as savings

Skydance's CEO told investors AI could matter more than CGI for original films. The shareholder letter filed the same day books AI only as cost cuts and merger speed, and he named no tool.

Skydance's chief made the indie AI filmmaker's pitch to Wall Street on Monday, then his company wrote AI down as something else. For makers hoping studios will fund cheaper original bets, the gap between what was said and what was filed is the story.

On Skydance's first earnings call since it bought Warner Bros. Discovery, David Ellison said AI "really has the potential to be the biggest creative unlock in modern history," bigger than the CGI wave of the early 1990s. His example was The Terminator: made for $4 million, and by his estimate "$60 million-$70 million" to shoot shot for shot today. Cheaper tools, he argued, make the risk of original stories "more digestible."

The shareholder letter filed with the SEC the same day is narrower. AI appears three times, all as savings: back-office AI inside an overhead cut of at least $2.2 billion, coding tools inside a $1.4 billion technology cut, and the reason Paramount+ and HBO Max can be folded together in about a year. Ellison said Skydance is developing AI tools but named none, and the AI question was picked in advance and read out before the live Q&A.

His own math needs a footnote as well. The Terminator was budgeted at $4 million but finished at about $6.4 million, and Skydance's two Terminator sequels cost $155 million and $185 million.

Our take: Ellison is right that AI lowers the price of a risky idea, and it matters that he says so out loud. Until a budget line or a named production carries that unlock, read it as a pitch to investors rather than an open door for filmmakers.

Source: Skydance Q3 2026 call transcript and shareholder letter (SEC Form 8-K), Nov 9, 2026.

Read the full article: thegenerated.com/stories/ellison-ai-creative-unlock-skydance-filing-20261110

An AI knockoff of El Empleo went viral, then sent viewers to a $49 monthly course

An Instagram account's AI remake of the award-winning short El Empleo came down after co-director Patricio Plaza objected. The reported 5 million views are unverified, and the account denies being paid by the course it promoted.

An AI remake of El Empleo, the award-winning 2008 Argentine short, reportedly hit 5 million views in under a day. Then its caption turned that audience into leads for a $49-a-month course, and every honest AI maker pays when viewers learn to read a viral AI post as a sales pitch.

The Instagram account Nightframe, with about 16,000 followers, posted a 30-second photoreal restaging of the hand-drawn short by Santiago Bou Grasso and Patricio Plaza, which won the FIPRESCI prize for short films at Annecy in 2009. The caption told viewers to send the word AI to learn how it was made. That opened an automated chat in Nightframe's own inbox, which asked people to follow the account before it sent a link to AI Film School, a paid online community.

Plaza objected and the video came down. He said the account was infringing the film's copyright to make an AI business go viral, which is his view of the law, not a ruling. The operator told Cartoon Brew they are a student running the page as a hobby, that the original was credited, and that the school has never paid them. We are not alleging a payment, and the 5 million figure is reported: no screenshot shows it.

Our take: crediting a film is not permission to remake it. A restaging of someone's whole short is an adaptation, and once the post sells a course it is marketing built on two animators' work. Ask before you adapt a short, and show how you work on original pieces.

Source: Cartoon Brew, Nov 9, 2026; Instagram; Skool.

Read the full article: thegenerated.com/stories/el-empleo-ai-knockoff-course-funnel-20261110

Three US states now require AI labels on ads, but viewers in our poll want every AI use disclosed

New York, California and Pennsylvania ask for an AI label mainly when AI could mislead, while 87.8% of 1,333 voters in our paid X poll want films and ads to always say. The poll is self-selected and has no demographics.

Grimes's Accelerando video credits at least three AI artists alongside its dancers, choreographer and makeup team. Asked what those artists actually did, the team defended their work without naming a shot or a tool, and that gap is what viewers now notice: a vague credit can read as something hidden, even when no rule asked for anything.

The rules are moving, slowly. Pennsylvania signed an AI advertising disclosure law on October 22, the third US state after New York and California, whose version was written by state Senator Angelique Ashby and takes effect January 1. All three ask for a label mainly when AI could mislead, such as a realistic synthetic person selling a product. Audiences set the bar higher. In our X poll, 87.8% of 1,333 voters said films and ads should always say when they used AI. That poll was boosted with paid promotion and has no demographics, but two larger surveys point the same way: 90% of 2,001 UK 16 to 24 year olds for the British Council, and 74% of about 1,000 US adults for Cint.

Disclosure has a cost too. Across 16 experiments, Michigan Ross researchers found an AI label cut ratings of identical writing by 6.2% on average. That is the argument for a specific credit over a bare label, because viewers dislike AI people far more than AI scenery, and saying which one you used tells them what they actually want to know.

Our take: disclosure is turning into a craft credit. Name the task AI did, what stayed human and who did the AI work, and the credit reads like a crew list instead of a confession.

Source: Pennsylvania SB 806; California SB 1050; our Generated Voices X poll, Oct 21 to 27, 2026; British Council; Cint; Michigan Ross; Business Insider.

Read the full article: thegenerated.com/stories/ai-disclosure-laws-vs-viewers-ai-credit-20261110