We’re kicking off 2026 with some fresh thoughts and insights directly from two award-winning actors, producers and filmmakers in Hollywood today.
Almost exactly a year since the first time it happened, Ben Affleck has gone viral again with his comments on AI in the film industry.
This is all thanks to a recent appearance on the Joe Rogan Podcast, chatting alongside long-time collaborator and industry veteran Matt Damon.
Given our position as experts on responsible AI use in film & TV production, we wanted to discuss his key points, see where we align, and hopefully give everyone some food-for thought for everyone in the industry as we enter the new year.
(If you have time to spare, scroll down to the bottom of this article to see the original clip in full)
Key Takeaways
AI as a technical tool
1. AI is like electricity, it’s all about how you use it
Ben says: AI should be viewed as a technical tool that will change production processes (some for the better and some for the worse) rather than a total replacement for filmmaking.
Our take: Totally agree
AI in itself is not a bad technology, it’s all about how it is created, commercialised and used by someone.
2. AI excels at tedious, expensive tasks
Ben says: The technology is best suited for “filling in all the places that are expensive and burdensome,” such as rendering graphics or changing background locations (e.g., making a studio look like the North Pole) without having to travel.
Our take: Yes, with caveats
VFX companies can struggle under time pressure when tasks like rotoscoping and matte painting are involved, which already leads to a lot of AI use in these areas. However one person’s “burdensome” task is another’s passion and craft, and to get production-quality results a skilled human eye is typically still needed, with or without AI.
No matter how far businesses intend to take AI adoption in the VFX process, we have to be careful to balance the need to innovate and reduce costs with the need to grow the skills of people in these teams, supporting a steady pipeline of experts to continue large-scale quality work in the future.
Limitations of AI Creativity
3. AI writing is inherently mediocre
Ben says: Generative AI (like ChatGPT) produces “shitty” creative work because, by design, it “regresses to the mean” or the average of its training data.
Our take: Not quite.
This is not fully technically accurate, and there are knobs to turn and levers to pull to say how much you want a model to “play-it-safe” vs take risks and think creatively. However I think the point is really about the need for human involvement, not fully creating ideas with AI. Many studies have shown that the best results come from humans and AI models working together, carefully and intentionally.
4. AI cannot create something good from scratch
Ben says: It is unlikely AI will ever write anything meaningful or create movies “from whole cloth,” though it may serve as a minor tool for writers to brainstorm specific logistical scenarios.
Our take: Not exactly
Trying to create a system that is fully AI-based (from idea to execution) is perfectly possible today, but it will not result in something impactful and moving. That said, there are so many valuable ways to use AI in the process of filmmaking that I feel this undersells it’s capabilities quite significantly.
Economics and “Hype” of AI
5. Fear of AI is driven by investment narratives
Ben says: The “existential dread” that AI will wipe out the industry in two years is a narrative pushed by tech companies to justify high valuations and massive capital expenditures to investors.
Our take: Agree.
We do need to take AI’s impact and potential seriously but all the singularity talk is not helpful and not realistic.
6. AI technological progress is plateauing
Ben says: the improvement curve for AI models is leveling off; newer models (like GPT-5) may be only incrementally better than previous versions but are significantly more expensive to operate regarding electricity and data.
Our take: Yes, but…
When it comes to more focused AI tools ( for example AI-driven scene relighting), there’s still lots of progress being made and more to come in the next few weeks, months and years.
Legal Protections and Guilds
7. Likeness rights are already protected from AI
Ben says: Current laws regarding name and likeness prevent companies from simply selling or using an actor’s image without permission.
Our take: Yes, but..
This simple fact is not enough to protect people in practice. For example, if you’re a well known face, productions can unintentionally create content using AI that resembles you without ever using images of you to do it. Others may try uploading protected assets of an actor into AI models in the early stage of a project, assuming it is low risk. Even though many actors prohibit this in their contracts now, many people working in the industry do not understand the real risks involved in these kinds of actions and mistakes happen often.
8. Unions will manage the transition of AI’s integration into industry
Matt & Ben say: Guilds will establish guidelines for AI use, similar to existing rules regarding the number of union extras required on set.
Our take: Yes, but it’s tricky
Many guild and union guidelines already exist. But the detail and complexity of these guidelines keeps increasing as the industry tries to carve out protected types of work for humans. It’s a challenge for production companies to keep on top of it all, and will only get harder as distributor requirements, insurance requirements, and changes in international AI leglislation continue.
Value of Human Artistry
9. The “Handmade” Premium
Ben says: As AI-generated content becomes ubiquitous, Affleck predicts audiences will place a higher value on “real things that are made by real people,” similar to the appreciation for handmade furniture.
Our take: Yes, but for how long?
We agree that there will always be interest in human-made works. But the real decider will be the cost to the audience of partaking in that media and experience. Human-made options may become prohibitively expensive to make, have larger budgets to recoup with a smaller audience, and have highly similar content to compete against which is mostly made using AI. In this case, the market for human-made features will either slowly dwindle or become a privilege only for the elites.
10. AI lacks lived experience, True art requires memory and trauma
Matt says: While AI can create photorealistic images, it cannot replicate the “complications of real life experiences” that allow an actor to convey genuine emotional depth. Take Dwayne Johnson drawing on the real trauma of his father’s addiction and his mother’s cancer diagnosis to perform a scene in the recently awarded “The Smashing Machine“.
Our take: Broadly, yes
Current models are trained on so much information from real humans that they can decently simulate some degree of human experience. But if the real value of a particular story to audiences is that it is about real person then using AI will always leave something lacking.
As always, if you need guidance on how AI actually applies to your production or VFX busienss, reach out to us at hello@aimici.co.uk
