At BearJam we’ve been working with AI in video production for a couple of years now. What started as a tool for small quick fixes has evolved into a full medium of its own. We’ve produced fully AI-generated ads and content, as well as plenty of hybrid projects too.
Over the past 18 months, AI video has got extremely good. The tools are genuinely capable. The costs are falling. The creative possibilities are expanding faster than most people in this industry have had time to process.
So why is adoption still so slow?
It isn’t the technology. The technology works. The barrier is trust, and until the industry gets serious about addressing it, we’ll keep having the same circular conversation about whether AI belongs in professional production.
It does. Here’s what’s in the way.
Why AI-Generated Video Still Has a Quality Problem
It’s the bad work that travels fastest. A strange AI hand or melting face gets shared as evidence of the technology’s limitations, but a quietly effective AI campaign centred in truth, insight or emotion rarely does. Confirmation bias is doing a lot of the work.
Yet audiences have always accepted constructed images, from heavily manipulated product photography to elaborate practical effects. We don’t usually judge creative work by whether it was made in-camera, digitally or through a complex production process. We judge whether it works.
The same standard should apply to AI. The tool doesn’t determine the quality. The thinking does.
AI Copyright, Creative Ownership and Consent
Who owns AI-generated content? Where did the training data come from, and who agreed to what?
This is the trust issue with the most legitimate foundations, and the one the industry is right to take seriously. The question of justice for creators whose work has been used to train models without consent hasn’t been resolved. Until it is, a layer of unease will remain, and that unease will slow decisions.
Logging prompts and proving provenance is a reasonable step. But if the underlying model is built on contested foundations, that only goes so far.
There’s also a bigger question worth sitting with. When I listen to Sam Fender on a streaming platform, he receives a royalty. Could there be a world where asking for creative work in the style of a specific artist means that artist receives something in return? I think that’s the direction of travel. The platforms that get ahead of it, rather than waiting to be regulated into it, will earn more trust from the creative community, and from the brands that depend on them.
Distrust of AI Platforms in Production
Underneath a lot of this sits a broader geopolitical unease. The leading AI video platforms are predominantly American or Chinese. For brands and broadcasters with data governance obligations, that raises real questions about where prompts are stored, how outputs are used, and what rights apply.
It isn’t paranoia. It’s due diligence. And the platforms haven’t always made it easy to find clear answers. Until there’s more transparency, and ideally more European alternatives in the mix, this will remain a genuine blocker, particularly for larger, more cautious organisations.
The platforms need to do better here. Trust isn’t granted; it’s built. And right now some of them are making it harder than it needs to be.
Why Human Creativity Still Matters in AI Video Production
Creatives are understandably wary. Professional identity is closely linked to the skills we’ve spent years developing; the eye, the instinct, the craft. When a tool arrives that appears to replicate some of that, it doesn’t just feel like disruption. It feels personal.
But the creatives I’ve seen thrive with AI aren’t the ones who ignored it. They’re the ones who used it to do things that weren’t previously possible on their budgets, timelines, or briefs. The skill shifts, but it doesn’t disappear. Prompt craft, creative direction, taste: these still matter enormously. Maybe more than ever, because the gap between a good AI output and a mediocre one is almost entirely down to the human behind it.
As for job security: if you want to stop your lunch being eaten, learn to cook differently. The roles most exposed are those built around executing repeatable tasks. The roles that are growing are built around judgment, relationships, and creative direction. The production industry has navigated this kind of disruption before; from film to digital, from broadcast to social. Each time, the industry adapted. This moment is bigger, but the logic is the same.
Fully AI-Generated, Hybrid or Traditional Video Production?
One of the most useful reframes I’ve found is thinking about AI involvement as a spectrum rather than a binary. At one end: fully AI-generated, concept to delivery. At the other: traditional production where AI plays no role. In the middle, and this is where most of the interesting work lives right now, is the hybrid: AI-assisted concepting, AI-generated elements composited with live action, AI handling versioning and localisation at scale.
There’s no hierarchy here. The right approach depends entirely on the brief, the budget, and what the work needs to do. Being transparent about where on that spectrum a project sits is itself an act of trust-building; with clients, with audiences, and with the wider industry.
Knowing which model fits which situation is the craft now.
How AI Is Expanding Access to High-Quality Video Production
I had a call recently with a prospective client, a start-up launching a new supplement. They needed video content covering multiple use cases, audience segments, and geographic regions. A traditional shoot to meet all those requirements was out of their budget. They’d had a quote from a company overseas at rock-bottom prices, but they wanted a partner who would understand their brand and be around to iterate. They were prepared to pay a bit more for that.
That conversation stuck with me. Because what they were describing wasn’t really a conversation about AI at all. It was a conversation about trust.
This is the opportunity that gets overlooked in the debate about whether AI is good or bad for the industry. AI opens doors to clients who previously couldn’t access serious production. It makes ambitious creative viable at budgets that would have ruled it out. And when those relationships are built well, when you become the partner who understands the brand, thinks strategically, and delivers work that performs, the brief grows. What starts as an AI-led content project becomes a deeper engagement. More ambitious work follows. Sometimes that means more AI. Sometimes it means a full live-action production. The tool is a starting point, not a ceiling.
Client expectations have moved. Many brands are now actively asking for AI in the pipeline; they want the scale, the speed, the creative ambition it unlocks. Agencies that can’t offer it will find those conversations going elsewhere. But the real edge won’t go to the agencies that simply adopt the tools. It will go to the ones that use them in a way that earns trust; with clients, with creative partners, and with audiences.
The expertise gap between those who’ve been genuinely working with this technology and those watching from the sidelines is significant. It won’t stay open forever.
AI has a PR problem. But it’s a solvable one. The agencies who do the work, build the trust, and bring their clients with them won’t have that problem for long.
About: BearJam is an award-winning London-based video production company combining traditional production expertise with AI-powered production.
