Paradoxical Truth: the consensus is that generative AI creates an existential crisis for Hollywood by commoditizing high-production video while Hollywood’s incentive structure prevents it from competing with AI – a classic innovator’s dilemma.
However, the variant perception is that AI actually solves Hollywood’s oldest economic bottleneck: the un-scalability of the human actor or celebrity. By decoupling the “celebrity persona” from biological constraints, AI will transition the industry from a high-risk, hit-driven box office model to a high-margin, recurring-revenue software model. The value does not vanish; it migrates from the physical production of content to the infinite licensing of digital IP by an every increasingly intelligent and “creative” AI producer.
Hollywood will undergo significant AI Metamorphosis
A traditional analysis of this subject matter would go into Hollywood’s storied history, the current ecosystem and dynamic of how TV and films are made, and the art of creative video production. The author of such a traditional analysis would someone from the “inside” – a Hollywood veteran, maybe a director, producer or well-informed actor/actress.
Well I would be wholly unqualified in such an approach because (1) I have no experience in professional video production or am I in the industry and (2) my subject matter expertise is technical in nature and largely in AI tech, not Hollywood. Rather, my approach here will purely be based on economics and the progress of AI development – both of which is well-modeled (economics) or very likely to happen (AI exponential evolution).
So let me cut to the chase and five predictions in light of this innovator’s dilemna.
1. In 10-15 years (if not sooner), there could be a 20/80 split between real movies and AI-generated movies (80% of movies will be AI-generated).
2. Due to economics (salaries, royalties, production costs), $NFLX and $AMZN will find it unprofitable (fewer impressions and high costs) to actually produce real movies at the scale they once did, or more importantly scale to the growing consumer demand.
3. The market for real-life movies will be the “indie” section of Netflix, not on the front reel. Yes, real-life movies will continue to exist for a while (before possibly being phase out completely in the distant future) because we want a sense of “human connection” – a sense of authenticity, of lived experience, but these are human fragilities/weaknesses that AI will try to capitalize on and will train on to mimic.
4. AI-generated movies will have evolved past the diffusion level of realism – now incorporating LWMs (large world models). With these LWMs, the physics, dynamics, motion of characters will be indistinguishable from real life. Photorealism and ray-tracing enhanced by AI estimation techniques will make image quality awe-inspiring.
5. Existing actors/actresses will morph into digital personas and their likeness (voice, appearance, personality) will be used as training for other digital personas. This is already happening with Matthew McConaughey and Michael Caine recently working with ElevenLabs to training their voice LLMs. Don’t get me wrong, the Taylor Swifts and Sydney Sweeneys will continue to exist – people have always enjoyed idolizing stars and heroes.
My Name is Hollywood, King of Kings
AI is clinically and ruthlessly rational and efficient – it doesn’t care if humans believe that certain aspects of themselves (like creativity) are irreplaceable. Humans are clinically and emotionally flawed – they attach themselves to an Ego who they think represents their true identity – an identity they believe is unique and irreplaceable. For an industry still in denial of its fungibility, this is the precise reason that AI will have an even more “unexpected” and devastating impact.
Jude Christensen 
Leave a comment