Welcome back to the AI For Humans newsletter!

Are you aware there’s a new Avatar movie coming out? Are you alive in the year 2025?

The Disney machine is in absolute overdrive for Avatar: Fire & Ash (my advice: see in IMAX and don’t wait for streaming) and, as part of the hard press, James Cameron himself is doing a ton of interviews for the movie.

He’s also been talking a lot about AI…

I am a massive James Cameron fan. If you were a kid of the 80s, 90s or 2000s, it’s hard not to be. Not only was he one of the most prolific & successful directors of his era, he embraced new technologies to push forward the entire medium.

What’s interesting about this particular headline is that it reflects one aspect of Cameron’s take on AI but doesn’t tell the whole picture.

Cameron, as you’ll see below, is actually quite bullish on generative AI as part of the production & creative pipelines in Hollywood.

Or at least… he was. Or he still is. It’s really hard to know.

In fact, it might be Cameron and others’ definition of Generative AI itself or lack thereof that’s in part driving the debate and confusion within Hollywood around an entirely new technology.

Let’s look quickly at what Cameon actually said on this CBS Sunday Morning piece that generated the headlines above (around 6:45 in this video).

The direct quote is below:

Love this newsletter? Forward it to one curious friend. They can join in one click.

Cameron has been speaking on this press tour about the connection he makes with the actors on the set of the Avatar films even though a lot of the production is handled via comptuers.

Specifically, he’s spoken about how new technologies (most of which are not Generative AI but def have AI embedded within them) have allowed for him to focus in more directly on the ‘human-director talks to human-actor’ part of the process.

I encourage you to listen to this two-part interview with him on The Town podcast to hear more about this (and a bunch more about AI as well).

HOWEVER…

Cameron is also, maybe contradictingly, quite bullish on Generative AI.

He’s on the board of Stability AI, the company behind the Stable Diffusion video model.

In this video recorded for the AI + Robotics Summit he says that he “plans to be on the leading edge of applying AI to my storytelling”.

And, in this April 2025 interview with Meta’s Chief Technology Officer Andrew Bosworth, he even says we shouldn’t be litigating AI’s inputs for the video models but rather their outputs. This might be considered, by some, a fairly radical take on AI media rights.

So what’s going on here? Does James Cameron hate AI or not?

It’s become harder and harder to hold two separate truths at one time.

The world is shouting at us constantly to ‘pick a side’ in every argument and rewards us online with clicks, likes and shares from those whose views we echo.

In this particular instance, it’s clear that Cameron likely draws some sort of line at re-creating acting performances with Generative AI while still being ok with the use of it in specific production pipelines or to make technical creative work much faster.

But, if the human acting input is the issue, where would he stand on something like Runway’s Act Two tool that AI creators like Neural Viz have used so successfully?

I personally assume that Cameron would see the work of Neural Viz and kind of be amazed at what was capable with off-the-shelf tech and one person working in their home.

But, if you read just the quotes he gave in the CBS Sunday Morning story, you might believe he’s 100% anti-AI and against everything it stands for.

I guess what I’m trying to get at here is the idea of how complicated this conversation gets when the nuance is removed. And it feels like something the Hollywood mainstream press is constantly doing with every Gen AI story they write.

James Cameron can 100% believe in Gen AI as a huge new tool for creatives while also saying he doesn’t want to use ‘text-to-video’ actors. That is ‘horrifying’ to him. It might be to you as well.

But, ultimately, he can hold both of these truths at one time.

And, I hope, so can you.

See you next week!

-Gavin

In this week’s AI For Humans: No new show! We took the week off for Thanksgiving.

But check out our Thanksgiving special from two years ago👇

Sometimes the smallest AI image prompts have a lot going on under the hood.

One of our favorite AI creators TomLikesRobots shared this longer Nano Banana Pro prompt to compare you to a number of different celebrities and see which ones you look the most like.

What’s fascinating with Google’s AI image model is that because it has the insanely powerful Gemini 3 running alongside, it can do all this reasoning work to allow you to find out that if you make a Zoolander-like face in a photo, you too can get Ben Stiller as one of your celebrity look-a-likes.