In April 2026, the Polish Association of Advertising Producers (SPR) published something the industry has been missing everywhere: not another AI opinion piece, but a structured measurement of what production companies, agencies and brands are actually doing with AI.
The report, “AI w produkcji reklamowej 2026,” combines a quantitative study run by IQS (136 respondents from production houses, agencies and client-side marketing teams), ten in-depth interviews with producers and post-production specialists, and fifteen written commentaries from executives including Nestlé Poland’s CMO, Havas Warsaw’s Head of Production, and Bites Production Hub’s own CEO Jakub Laskus, who contributed the report’s chapter on food content specifically. SPR calls it the first study of its kind in Europe.
The dataset is Polish. The fault lines it maps are not. Production houses everywhere are running the same experiment: how much AI to let into a workflow, who pays for the shift, and who is still liable when it goes wrong. Six findings stood out.
Adoption has outrun strategy by a wide margin. 76% of Polish production companies delivered at least one AI-assisted project in the past twelve months. Only 13% did so under a documented AI policy or plan. For the remaining 87%, AI usage means individual projects, informal client agreements and no internal audit trail. The report’s warning is blunt: the industry is adopting the technology faster than it is building the decision-making infrastructure around it, and that gap is where legal, financial and organizational risk accumulates.
Hybrid is not a phase. It is the market. Three production modes now coexist rather than replace each other: classic (roughly 34%), hybrid (53%), and full AI (13%). Respondents were explicit that this is not a queue where everyone eventually lands on “full AI.” A producer interviewed for the report put it plainly:
“We work traditionally, hybrid and AI. Each path has its own clients, budgets and risks. It’s not about which one wins. It’s about which one makes sense for a given brief.” For hero campaigns specifically, 88% of respondents said AI functions as a supplementary tool, not a replacement for camera-based production.
Trust collapses the moment a human face appears. This is the sharpest number in the report. 79% of respondents accept AI-generated backgrounds and environments. Acceptance for inanimate objects sits at 64%. For AI-generated actors, meaning face, body or voice, acceptance drops to 16%, a fivefold difference from backgrounds. The split tracks risk exposure just as clearly: AI acceptance for tactical digital and social content is 77%, but for brand-defining hero and TV campaigns it falls to 23%. Jakub Laskus’s commentary on food specifically noted that despite 76% of respondents having already worked with AI, acceptability in image-driven brand campaigns runs far lower, and that in food and beverage, authenticity and consumer trust matter more to brand owners than time or cost savings. His conclusion: camera-based food production will hold on longer than the current industry conversation assumes, not because of budget size, but because of how close the content sits to the brand’s promise.
AI moves cost around; it does not delete it. 75% of production companies still price AI work using the old model: hours worked and technical complexity. Meanwhile, the report’s interviews describe AI reducing cost in pre-production and shooting while increasing it in producer oversight, correction, quality control and rights management.
61% of respondents cite “AI look” and insufficient control over detail as one of their biggest quality challenges, precisely the problem that generates extra post-production hours.
One post-production specialist quoted in the report: “Because the weight of the shoot shifts onto AI, post is not automatically cheaper. It can be more expensive.” A network agency producer, describing a direct AI-versus-3D cost comparison on one project: “The difference wasn’t nearly as big as we assumed. We talked the client into going traditional after all.”
The law is not the problem. Not knowing how to use it is. 92% of respondents name legal barriers as their top constraint on AI adoption, more than any other factor, including budget (cited by just 9%). But IP lawyer Jan Wiegner’s commentary in the report pushes back directly on the premise: Polish copyright law, the civil code and unfair competition law already cover most of the situations production teams encounter. What is missing is not regulation but market practice: knowledge of the rules, contractual standards, and above all, protection against the one-sided licensing terms that AI tool vendors are currently writing into their terms of service. Broken down, the specific legal worries respondents flagged were: lack of clear regulation (59%), risk of infringing someone else’s copyright (51%), no clear rights to AI-generated outputs (49%), no contractual standards in the industry (49%), image rights exposure (45%), training-data risk (43%), vendor licensing restrictions (40%) and unclear liability (37%). That is not one problem. It is eight, all unresolved at once.
The producer’s job is not shrinking. It is absorbing more. Of all six findings, this is the one the report says achieved full consensus across every source: quantitative data, interviews and expert commentary alike.
AI does not redefine what a producer does. It expands the list of things a producer has to know, stacking creative, technological, financial and legal judgment onto a role that already carried all four.
New specializations may emerge around it (the report floats the idea of an “AI operator” role embedded in production teams), but they function as an addition to the producer’s coordination, not a replacement for it.
SPR closes the report with three concrete asks for the industry rather than predictions: model AI-ready contract clauses covering copyright, chain of rights and vendor liability; shared legal and competency education across clients, agencies and producers rather than each party learning in isolation; and a common pricing vocabulary for classic, hybrid and full-AI production, since hours and technical complexity no longer describe what teams are actually being paid for.
If you want to dive deeper into the report, you can download it here.
