I’m still trying to work out what I actually want AI to do for me as a creative, and that sounds odd given how much of the technology we’ve all been looking at over the past couple of years, but I think that uncertainty is probably useful. I can see the appeal of having a machine chew through the boring stuff; I can see the potential for getting from an idea to something tangible much faster, but I’m not particularly interested in a future where the answer to every creative problem is “let the AI do it”.
Which is why Autodesk’s big AI pitch at AU26, held this year in Las Vegas, has caught my attention. The company is talking about “closing the gap between what we can imagine and what we can build”, which is corporate language, certainly, but there’s something behind it that I find more interesting than the usual AI promise of making things faster, cheaper and with fewer people.
Raji Arasu, Chief Technology Officer and EVP at Autodesk, put the idea more bluntly during the keynote: “Our mission [is] to unlock capacity for you.”
More than any individual AI feature, this is what I think Autodesk is actually selling at AU26. The company is essentially arguing that AI should increase what people can accomplish with the resources they already have, whether that’s a designer trying to get through a mountain of repetitive work, an engineer juggling increasingly complicated product data or an animator trying to get a character to do what they imagined in their head without spending half the day fighting keyframes.
Can AI take away the drudgery?
After years of speaking with animators and game artists, that last one is the easiest for me to wrap my head around. Anyone who has spent time animating, or in any creative job, knows there’s the fun stuff and then there’s the stuff you have to do because, well, that’s how animation works. If AI can take some of the second category away while leaving artists in charge of the first, I’m interested. If it starts making the creative decisions for artists and animators, I’m much less interested.
That distinction runs through much of what Autodesk is showing at AU26, across architecture, engineering, manufacturing, and media and entertainment, and points to a general direction of travel for digital and 3D arts. Take the new standalone, agent-first version of Autodesk Assistant. Instead of making you figure out which Autodesk application contains the tool you need, the idea is that you start with the outcome you want, and Assistant uses your project context to determine which Autodesk capabilities are relevant.
I like that direction, partly because I’ve spent enough time with creative software myself, even just as a hobbyist and journalist and not a pro, to know that finding the right command, tool, or workflow can sometimes be half the battle. More importantly, though, Autodesk is trying to make the AI understand the work you’re doing rather than simply understand a prompt and do it for you.
Arasu described that ambition as “understanding not just what’s happening on your project, but why and what to do next.”
It's a small distinction, but it shapes how you use and approach AI; here at Autodesk, a creative project isn’t a blank box waiting for you to type something into it. It has models, drawings, materials, revisions, decisions, mistakes, constraints and a fairly sizeable pile of accumulated context. Autodesk wants its AI to work with that stuff, which is why the company keeps coming back to the slightly grand term “project intelligence”.
In less corporate language, I think the idea is that the software should actually know what you’re doing, help you do it, but not take over. Arasu put it another way when she described how, once AI understands a project’s context, “it becomes grounded in your workflow, your company, and your project ecosystem”.
The broader AI take-up
When you get into Autodesk’s creative tools, that subtle shift from replacer to helper piques my interest. Flow Studio is probably the AU26 announcement and demo I’m most curious about, coming tomorrow. The idea is that creators can create and assemble characters and AI-generated worlds, then bring them into an editable 3D environment to direct and refine what they’ve made.
We’ve already seen a lot of AI demos where someone types a sentence and gets a character, a landscape, a shot or an entire little film spat back at them, and yes, it’s impressive, but I’m starting to think the generation is actually the least interesting part of using AI creatively.
Give an animator a character in a blank space and tell them they can now decide how that asset moves, how they’re framed, what happens to them, what needs changing and what needs throwing away, and animators will be back in territory that feels much more like making something and a lot less like prompt jockeying to a final scene. It’s the difference between generation and direction, and I suspect a lot of the genuinely interesting creative work around AI will happen in direction.
I've already covered why MotionMaker in Maya fits into this approach to how AI can and should be implemented. Autodesk uses machine learning to reduce manual keyframing in character animation while keeping the artist in control of the final performance.
Again, I don’t particularly care that a machine can produce movement; I care more about whether it can save artists and animators from the parts of the process that keep them from spending time on the performance itself. If AI can get a rough motion in place quickly and then an artist can spend time pushing the pose, timing, weight and personality around until it feels right, that’s a compelling use of AI.
The machine has done some work, but it hasn’t really made the creative decision for you. And that, I think, is where Autodesk’s wider AI strategy starts to make more sense, even if you have to wade through a lot of Autodesk terminology to get there. Across the company’s different industry clouds, the pitch is broadly the same: give people more useful information, remove repetitive work and let them spend more time making decisions that actually matter.
Arasu summed up Autodesk’s intended outcome at the end of her section of the keynote: “less searching, less stitching, less AI complexity, more creating, more solving, and more innovating.”
And while it feels like a healthier AI place to be than simply saying “look how much content we can generate now”, it also leaves the bigger question hanging: if AI increases the amount of work one person can do, what happens to the number of people a company decides it needs?
The impact of AI won't go away
Autodesk can talk about unlocking capacity, helping businesses win more work, and making projects more predictable, but that doesn’t magically remove AI's economic reality. A tool that lets one animator do the work that previously took two people could be fantastic for the animator who gets to spend more time on the creative side of the job, but less fantastic for the second animator who no longer gets the job.
I don’t think Autodesk has an answer to that, even though CEO Andrew Anagnost has been honest about the impact on jobs, and I don’t think any software company can simply declare the problem solved. What I do think is useful is that the company is at least making a case for AI as an extra layer around human expertise in all creative sectors, rather than presenting the human as the inconvenient bit of the workflow.
That’s why the Flow Studio developments and what Autodesk shows around MotionMaker interest me more than another shiny generative demo. I want to see what happens when AI gets good at the tedious parts and leaves the messy, subjective, frustrating, very human business of deciding what something should actually be like to the person making it.
Maybe that’s where the real test for creative AI will be, not whether it can make something impressive from nothing – we already know it can – but whether it can give an artist enough extra time, control and freedom to make something better than they could have made before. That’s a much harder thing to demonstrate in a keynote, and it’s also the bit I’ll be watching most closely at AU26.
Visit AU26 for more news and announcements as they happen.
