Takeaways from the GenMedia Conference on AI, craft, and why generative media might mean more original ideas and fewer sequels.
At this year’s GenMedia Conference, one of the more interesting questions wasn’t what AI can make. It was what we might finally be able to make because of it.
Which is a much more interesting question than whether a chatbot can write a screenplay. Although, to be fair, plenty of humans have been doing that badly for years.
A few things stuck with us.
The tools change. The argument stays the same.
Sean Bailey, former president of production at Disney, compared the current AI conversation to the arrival of CGI. Hollywood has always been a technology industry, he argued, and new tools have a habit of reopening old arguments about craft.
Every technological shift seems to come with a prediction that creativity is about to die. Creativity, inconveniently, keeps turning up to work.
But there’s a real question underneath all the noise: when does a new tool expand creative possibility, and when does it start undermining the people who make the work worth watching?
Nobody’s praising the camera.
That was the title of a session with creative technology veterans Stephan Bugaj and Jeffrey Kember, reflecting on what the transition to CGI can teach us about generative media.
One idea stood out: art directability matters.
Bugaj spoke about the level of control artists need over even the smallest details, from a character’s glasses to the texture of a garment. The point isn’t simply that AI can generate an image. It’s whether a creative can get that image to express exactly what they intended.
The technology can make things. The artist still needs to know what the thing should be.
The slop machine is not the dream.
Speaking about Amazon MGM Studios’ AI Studios, Albert Cheng pushed back on the idea that AI filmmaking means typing a prompt and watching a movie appear.
He described a process involving hundreds of people, from storytellers to VFX artists, whose job is to direct, refine and curate the work. Not exactly one person pressing a button and heading off to lunch.
The point is that human judgment still matters. Someone has to notice that the pixel is wrong. Someone has to decide the scene isn’t working. Someone has to know when to stop.
That last one may prove particularly important.
Of course, the creative possibilities don’t make the ethical questions disappear. Consent, ownership, training data and what happens to the people whose work helped make these tools possible remain very real concerns.
You can be excited about the technology and uncomfortable with parts of the industry building it. These are not mutually exclusive personality traits.
What if we made fewer sequels?
This might have been our favorite thought of the day.
Kember pointed to the economics of entertainment: making things cheaper and faster could allow individual artists to attempt projects that would previously have been too expensive, too risky or too difficult to get off the ground.
It’s not hard to see why Hollywood keeps making sequels. Familiar characters and established franchises come with built-in audiences, making them a safer bet than an entirely new idea.
The result? A lot of stories we’ve seen before, often with a number at the end of the title.
Kember’s argument was that generative media could change those economics. If ambitious projects become cheaper to make, more original ideas might get a shot. Filmmakers without studio budgets or industry connections could tell stories that would otherwise never make it past the pitch deck. More independent voices. Stranger ideas. Fewer boardroom-approved sequels.
The new original.
If the tools become widely available, access to the tools alone won’t be much of a differentiator.
Taste, intention, discernment and the ability to make something mean something will matter even more. As will the ability to tell the difference between a genuinely interesting idea and something that merely looks impressive in a demo.
Because the interesting question isn’t whether AI can make a film, an image or a campaign. It’s whether we use it to make more of the same, faster, or to make things we couldn’t have made before.
One outcome gives us an almost incomprehensible amount of content. The other might give us something worth paying attention to.
We’ll take more of the latter.
And if it means fewer sequels, even better.
(Although we will make an exception for Practical Magic 2.)
