By 2035 the “film industry” will still involve people on sets pointing cameras at actors, but much of what we call filmmaking today will have been permeated, and in places hollowed out, by machine-generated media.

Three of the most visible areas you mention — deepfakes, synthetic performers and machine‑written stories — are on their own trajectories that intersect in ways that change both the business and the culture:

• Deepfakes become a standard production tool, not a novelty. By the mid‑2030s, photorealistic face/voice replacement will be inexpensive, fast and legally normalized. Studios will routinely de‑age or age up performers, dub films into dozens of languages with perfect lip‑sync, and stitch coverage from multiple takes without anyone “missing their mark”. Continuity fixes and reshoots will often happen in post with an AI that can match an actor’s performance from hours of scanned reference. You’ll see classic IP extended indefinitely with the original star’s likeness licensed from their estate, and actors who are still alive renting out their face and voice for projects they never step onto a set for.

  The upside is creative freedom and cost savings; the downside is the erosion of trust and the need for new contracts, credits and watermarks so that audiences know what is authentic. Expect guilds and regulators to have pushed through “right of publicity” and consent laws that make unauthorized deepfake use a liability in legitimate production, even as illicit uses continue to proliferate elsewhere.

• AI actors fill in the spectrum between background and star. By 2035 you won’t just have digital extras and CG creatures; you’ll have wholly synthetic actors whose “careers” and personas are managed like K‑pop idols. Some will be bespoke to a film, others will be persistent characters with their own social feeds and fanbases. These AI performers will be used to reduce costs on mid‑budget projects, to do dangerous stunts virtually, to localize films to different markets (the Chinese version featuring a Chinese AI co‑lead), and to offer interactive, personalized experiences in VR/AR. They’ll also serve as a licensing revenue stream for human performers: a 70‑year‑old actor could “appear” in a teen drama through their AI avatar, with their consent.

  Unions will likely have negotiated provisions around “digital doubles” and minimum compensation for use of one’s likeness. There will be ethical debates about resurrecting deceased actors for new work and about the emotional authenticity of a romance performed by two neural networks. But audiences will acclimate to some synthetic presence on screen, just as they did to CG.

• Script generation automates a lot of the writing grunt work. Language models a decade from now will be able to outline, draft and iterate on full feature scripts in the style of any genre or franchise, and to take network notes as inputs. Development executives will use AI to generate dozens of pitches and treatments; writer‑directors will use it to break story problems and simulate how different scenes play. For commercial content, entire low‑end features and series episodes will be generated from prompts, with human oversight focused on compliance, brand fit and punch‑up.

  At the high end, human writers will still matter for taste, subtext and cultural acuity, but they’ll be working with AI collaborators rather than blank pages. One new market segment will be “personalized cinema” — films that are tailored to an individual viewer’s preferences, cast with their favorite actors’ licensed avatars and localized to their sensibilities, generated on the fly for streaming platforms.

These technical shifts will ripple through the industry structure:

• Production pipelines will be virtualized. LED “volume” stages, real‑time engines and AI‑assisted previsualization will shrink physical production footprints. An indie filmmaker will be able to generate a passable sci‑fi epic with a handful of people and a bank of GPUs, blurring the line between professional and amateur content.

• Business models will fragment. There will still be billion‑dollar theatrical tentpoles where practical stunts and marquee human stars are a differentiator. But there will also be an explosion of niche, on‑demand AI‑generated films and interactive stories on streaming services and social platforms. The long tail of user‑generated and semi‑generated content will compete for attention with studio output.

• Law and culture will adapt. Expect visible disclosure marks for AI‑generated imagery, new residual structures for use of one’s likeness and work in training data, and periodic scandals when a synthetic performance crosses a cultural or ethical line. The value of “authenticity” may become a selling point: films marketed as “no AI actors, no deepfake shots” in the way some music is marketed as “analog”.

By 2035, then, artificial intelligence won’t have replaced filmmaking so much as absorbed parts of it. It will lower barriers to entry, allowing more stories to be told and more experimentation. It will also commoditize certain kinds of content, forcing creative professionals to redefine their value. The films that endure will likely be those that use the tools to amplify a human point of view, rather than to simulate one.