What productions and agencies really worry about with AI in video

Almost every conversation about AI in the audiovisual industry starts from the wrong question. While the sector debates whether the machine is going to replace you, the decisions you actually have to sign off this quarter are stalling for four other reasons.

"Is this going to replace me?" is the fear that fills headlines, hallway chats and LinkedIn threads. And it is precisely the one that should cost you the least sleep. This article names the four real fears plainly and links, in each case, to the analysis that goes into detail. If you have to commission AI audiovisual production and you don't know where to start worrying, start here.

The wrong fear: "AI is going to replace me"

2026 has buried the fascination phase. The line that dominated Cannes Lions this year sums it up: the era of hype is over, now you have to prove it. Generative AI has stopped being an award-worthy trick and become infrastructure. The way Photoshop stopped being news and turned into the floor everyone works on, AI tools are becoming invisible. And once a technology turns invisible, what tells one supplier from another stops being "do they use AI?" and becomes "what do they do with it?".

There is a nuance that changes the mental frame. Recent research on creativity suggests generative AI doesn't flatten talent, it amplifies it. Work by Luchini and colleagues points to these tools widening the gap between those with craft and those without. Translated into your business: AI doesn't turn just anyone into a good storyteller. It turns a good storyteller into a faster one, and leaves the bad one just as bad, only producing more. The question that matters is a different one: does your judgment show enough to make you worth hiring?

The four real fears that actually stall a signature

Once you set aside the noise around replacement, four concrete doubts remain. Each has its own article.

• "Will this pay off?" The fear of paying for an efficiency that never turns into return. It's the question most repeated in a boardroom, and the one most suppliers answer worst. The real ROI, no smoke.
• "Which tool do I choose without getting locked in?" The worry of investing in the wrong model, depending on a single vendor, or uploading sensitive assets to an unclear platform. How to choose your stack without getting it wrong.
• "Will I lose my creative signature?" The fear that AI homogenises everything and your work becomes indistinguishable from the competition's. AI with human direction.
• "Can I scale without diluting my brand?" The doubt of anyone who needs to produce far more, more variants and more languages, without slipping into generic content. Scale without breaking the brand.

The risk people actually talk about: "AI slop"

There's a fifth factor that works differently. It doesn't stall the contract: it stalls the campaign. It's the public's rejection of badly used AI, what the industry now calls AI slop, and it's why many marketing directors slam on the brakes.

The data is uncomfortable. A Tracksuit survey of more than 6,000 consumers, reported by Business Insider, placed sentiment toward AI-generated ads firmly in negative territory: roughly 39% rejection against barely 18% positive rating. That's the difference between a campaign that connects and one that becomes a viral joke. McDonald's Christmas ad in the Netherlands and Coca-Cola's pieces were treated by much of the public as "soulless", and the brand ended up pulling content.

The consequence is almost paradoxical. Some brands have turned the rejection of AI into a selling point: Equinox, Aerie, Almond Breeze, earlier Dove and BMW, have run campaigns that boast about not using AI or that mock synthetic content. When your competitor brags about doing things by hand, the risk of badly applied AI damaging your brand stops being theoretical.

That's why at EficiencIAl we start from a simple idea: the fear of the person hiring is reputational before it is technical. Nobody fears an abstract fine; they fear the online pile-on with your client's logo on top. Doing AI well comes down to judgment more than technical capacity.

What's mature and what's still smoke

Being honest about the limits is, in this market, a competitive advantage. Here's our reading.

Status	What it covers
Mature today	Preproduction, previs and storyboard; editing and post (audio clean-up, colour, clipping); localisation and dubbing; and variant production at scale for advertising. On every one of these fronts AI already delivers measurable return under human direction.
Still smoke	Premium film or branded content generated autonomously, without strong direction. McKinsey is clear: AI-generated output still doesn't reliably meet the premium quality bar in film and television. Anyone selling you otherwise is selling a demo, not a process.

Our position: AI with judgment, not AI as a fad

We're not an AI company that has discovered filmmaking. We're audiovisual operators (shooting, drone, post) who have brought AI in where it genuinely adds value. That difference matters, because it means we know when not to use it.

Our edge is the fusion of real footage, drone capture and AI generation, directed by human judgment from start to finish. It's what pure AI can't offer and what protects your brand from the risk we've just described.

Sources

• Cannes Lions 2026: Publicis Groupe coverage and trade press.
• Tracksuit (consumer sentiment toward AI ads), via Business Insider.
• ADWEEK: ad campaigns boasting about not using AI.
• McKinsey: state of AI in media and entertainment.
• Luchini et al.: research on creativity and generative AI.

We publish figures with their origin in plain sight and without rounding them in our favour. If a data point expires or gets qualified, we correct it: it's the only way to hold a no-smoke position.

If you have to decide how AI enters your next production and you want to do it by judgment, not by fashion, let's talk. Start with the four analyses linked above, then move on to a conversation with numbers on the table.