This year has been quite a journey so far.
I just finished speaking as a panelist on the recent Social Peta Forum which was titled "AI Short Drama 2.0: From Content Generation to Content Intelligence". (https://socialpeta.com/en/webinars)
I was the first Australian to be invited to speak on their forum, which was very exciting, but my first though was whether I was even qualified enough to speak on this topic. But as it turns out, I was. It was just from a different perspective - from the point of view of a small indie filmmaker, based in Australia, where AI is still very much a hot topic (and not all positive).
I didn't set out to become an AI filmmaker. I come from a traditional film background, working on both sides of the camera. So my AI journey started, like a lot of people, out of curiosity — could I actually tell stories using AI?
What I found surprised me: it was still a genuinely creative process. I was building characters the way I saw them, setting up shots, editing everything in Adobe Premiere, adding music and sound design, just like I would on any live-action short I've produced and edited.
That's where most conversations about AI filmmaking stop — at generation. What can the tool make. How fast, how realistic, how much. But my own journey has moved me somewhere else entirely: from asking what AI can make me, to asking what I should make, and what I'm responsible for once I've made it. That shift — from content generation to content intelligence — is the thing I actually want to talk about.
Where the shift happened for me
It came into focus while I was exploring AI studios to make some episodes from one of my microdrama screenplays. These episodes were more realistic than those I had created for my first AI film, The Tell Tale Heart. I had created around 6 episodes, and had used my own actors headshots to create the main character, and the other characters where created based off my own vision of how I wanted those characters to look.
But whilst generating the videos, the AI studio started flagging them. It was telling me that the person in the video resembled some well-known person (who's name escapes me right now) - but the character it was flagging was me! It was my character, that looked like me, based off my own photo! And that's when it hit me — what if the other generated faces of the character's I created resemble someone real? I knew the flagged image, was me, but I couldn't say for sure that the other characters didn't look like someone. After all, don't we all apparently have a doppelganger out there anyway?
I didn't want to guess my way through that question, so I did the research and spoke to a former lecturer of mine, Julia, who now runs workshops on ethical AI generation through her business, Margarita Media. I know I used my own headshot create my main character, and built the rest of the characters from my own descriptions — but I've become conscious of that question ever since: could this look too close to someone real.
That's the actual dividing line, as far as I can tell:
- Generation asks what can I make.
- Intelligence asks what should I make — and what happens after I've made it.
AI slop is what happens when that second question doesn't get asked
I think a lot of the poor-quality, AI content we are seeing flooding social feeds right now comes down to that gap. Nobody stopped to ask whether a face was too close to someone real; or whether a story was worth telling; or whether a shot actually served the scene.
What you get instead is volume without judgement — content being generated with no intelligence attached.
What intelligence looks like in practice
For me, it's a pipeline, the same way I would treat any film production: start with the screenplay; break it down; create a storyboard; create the videos; then edit it myself. It's repeatable. I can run that same process on the next project, and I do. But the difference is that with AI, I'm also making sure I'm being ethical and responsible for my choices and my generations.
Circle of Secrets — The Curse of Evelyn is probably the clearest example I can point to. The story started as a stage play I wrote back in 1997 (you can read more about that in my last post). After several drafts, I finally re-wrote it as a traditional screenplay in 2021. And now, it's finally been re-created as a 20 episode AI-assisted animated microdrama - screening on the Anamana app, made through the Anamana 100 Creator Incubator (where I have been the only Australian to be selected so far - another first). So after nearly thirty years, it was modified in three different formats.
AI shaped this adaptation in two real ways:
The first was access — the original live-action version needed three filming locations, two of which required travel and accommodation for the cast and crew. That's the kind of cost that keeps stories like this in a drawer for a long time. I'm not rich. I'm just an indie filmmaker who has never been eligible for funding (apparently not successful enough to be considered), and for whom crowdfunding or self-funding wasn't always realistic in today's economical climate. By producing this story in AI, I completed it all in around four weeks, without taking much time off at all in my day job.
The second was form — The choice to make it animated rather than photorealistic wasn't a limitation, it was a decision, made for the same reason as the headshot: it let me tell this story without any risk of generated characters resembling real people (although I still created the main character based on myself, as I had always written it as a character I wanted to perform as an actor - but that was never going to happen now).
But what stayed entirely mine in this process were the way the characters looked; the details of the locations; the writing from a stage play being restructured into an episodic microdrama; every edited detail; every choice of music and sound design; and every judgement call about what to generate - what to use and what to discard. AI changed what was possible for this story. It didn't change who was writing and producing it.
Where the future is for AI filmmaking is heading
The part of this I'm most genuinely excited about is likeness licensing — actors, and even non-actors, being able to license their image for AI productions. I've been looking into this myself because I've been acting for over 30 years, but after relocating to Tasmania, (and getting older) the acting opportunities just aren't there anymore. I also know people who've always wanted to act but never had the opportunity or the ability to, and former actor friends who can't work anymore for health or personal reasons. AI means they can still be part of a story — and get paid for their likeness doing it. To me that's content intelligence applied at an industry level: not just "can I generate this", but "how do we build a production model where everyone in the frame actually consented to be there".
Tools generate content - but Intelligence is the discipline of deciding what's worth generating, doing it responsibly, and getting better at it every time.
I moved from one to the other fairly quickly in my own work — but I don't think that's true across the board, and I think our industry needs to be having this conversation a lot more than it currently is. Not about how fast we can produce, but what we're actually responsible for once we do.
Comments
Post a Comment