A Horror Film Slipped AI-Generated Frames Into Theaters Without Saying So
What happened
The horror film "Late Night with the Devil" arrived in March 2024 as a mock found-footage production set inside a 1970s late-night television studio. Three of its interstitial images, the brief still frames that separate segments within the fictional broadcast, had been generated using AI tools. The filmmakers did not disclose this before the film reached theaters. When viewers identified the frames online, the backlash was immediate.
Audiences who recognized the artifacts of machine-generated imagery flagged the images publicly, and the production company confirmed they had been produced with AI software as part of an effort to achieve the film's period aesthetic. Distributor IFC Films and streaming platform Shudder faced calls for a boycott from fans and working artists. Horror filmmaker Mike Flanagan stepped in with a public defense of the film and its makers, but that defense pulled more attention to the choice rather than settling the debate. Critics who had praised the film's craft before the disclosure found themselves revisiting what the word meant when applied to a production that had substituted machine output for commissioned illustration work without mentioning it.
The backlash was not primarily about the visual quality of the images. By most accounts the frames were competent. The objection was about the omission itself: audiences, reviewers, and industry workers whose income depends on illustration and concept art commissions had consumed something presented as creative professional work without knowing a machine had produced part of it. The film's commitment to period authenticity made the silence especially pointed. The 1970s aesthetic that gave the film much of its critical appeal was partly simulated in a way the audience had not been informed of.
The employment concern ran alongside the transparency one. AI image tools used inside a production pipeline displace the artists who would otherwise receive that work. When that displacement happens without disclosure, no mechanism exists for audiences, distributors, or working artists to make informed decisions about what they support financially. The incident joined a growing list of productions where AI substitution became visible only after public exposure, not before release, with no industry-wide standard requiring the gap to be filled.
What the controversy exposed is the absence of any mandatory documentation layer between production decisions and the audience that consumes them. Nobody required the filmmakers to record which images were machine-generated, who approved the decision, or how that choice was weighed against hiring a human illustrator. A provable record of what a system produced and when would not have prevented the creative decision, but it would have made quiet omission structurally harder and given audiences the information they were seeking before the film left theaters rather than weeks after critics had already filed their reviews.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.