AI comes up in almost every conversation we have with clients, and the tools change so fast that any fixed position dates within weeks. So this isn’t a rulebook. It’s where we stand today on using AI with care, and why the craft still comes first.
“The best moments in our films are the ones nobody planned: a laugh, a pause, a voice that catches.”
First, what counts as AI?
The label now covers almost everything. Editing software has cleaned up noisy audio for years using clever algorithms, and nobody called it artificial intelligence. Now the same kind of tool ships with an AI badge, because the word sells.
Or sometimes it doesn’t. A Washington State University study found that mentioning AI in a product description made people less likely to buy, because it lowered their emotional trust. The label is doing a lot of work in both directions.
We find it more useful to split the tools in two. Some repair and assist: rescuing dialogue from a noisy location, transcribing interviews or tidying an image. Others generate something new, like words, voices or pictures. The first group is simply good practice. The second is where we slow down and think.
How we use it, this week.
Honestly, a lot. It helps us shape and test scripts, draft voiceovers so clients can hear a script early, modify images where a project needs it, plan shoots and clean up audio recorded in places that were never meant for filming. It even helped us build this website.
A Harvard and BCG study of consultants captures what that feels like. On tasks inside what the researchers called a “jagged frontier”, people using AI worked about 25% faster and produced noticeably better work. On tasks just outside it, they were less likely to get the right answer than people working without AI at all.
That matches our experience. AI can be brilliant at one job and confidently wrong at the next, and the line between the two moves every few weeks. Knowing where it sits today is part of the job now.
People can tell. For now.
Merriam-Webster named “slop” its word of the year for 2025, defining it as low-quality digital content produced in quantity by AI. Audiences have learned to spot it: the glossy sheen, the missing grain, the voice with no grit in it. Once someone notices, they start doubting everything else on screen.
Spotting it is getting harder, though. A 2025 study from Queen Mary University of London found that listeners could no longer reliably tell cloned voices from real ones. So we can’t rely on the audience to catch it. The responsibility sits with the people making the film.
Respect for the people behind the work.
AI hasn’t changed how much we value the people we work with. Voice artists, composers, illustrators and freelance crews have been part of our best work for years. Used well, AI often frees up budget for the moments that really need them.
So we use AI as a tool, with care. Where a project needs a real voice, a real performance or an artist’s hand, we bring one in. Where a budget calls for something leaner, we explain plainly what AI can and can’t give, and let the client decide with their eyes open.
Staying nimble without cutting corners.
Our clients are rightly budget-conscious, and they want speed. Ignoring these tools would mean charging them for time we no longer need to spend. Using them carelessly would cost them the quality they came to us for. The job is to sit between the two and keep moving.
Some of this piece will probably be out of date within months, and that’s fine. What makes a film work hasn’t changed: understanding the people in it and the people watching it.
What it means for you.
When you commission video, ask your production partner how they use AI. A good answer is specific about where it helps, honest about where it doesn’t and clear that someone experienced is still making the calls.