Opinion piece by Ricardo Campos: AI in creative industries

It no longer makes sense to say that AI is revolutionising the design and creative industries. Many professions, as we know them, are undergoing a very aggressive adaptation.

Nor does it make sense to spend time repeating what we’ve read on blogs, or heard on podcasts or Instagram reels. The rules of the game have already been set, and they were created by a few — but very big — players.

The good news is that easy access to technology allows us to take action and make decisions with (almost) total independence. Historically, in times of such drastic change, this opportunity had never been available — and we must know how to seize it.

Those of us working in this sector are doubly fortunate: we have artistic talent, and we also possess the gift of creativity. We also carry with us the spirit of sacrifice and constant learning that goes hand in hand with any profile linked to art, graphic design, audiovisuals, copywriting, advertising, marketing, and so on.

If we follow what the so-called gurus say on social media, what we receive is a bombardment of options, apps, and fears. We already know that the goal of these messages is to sell, and to instil fear (which means selling even more). That is why I am writing this article: to ensure we move forward in this scenario with strategy and control.

The first point I want to share is that, according to the latest MIT report State of AI in Business 2025, which I analyse in detail in this neurona article, only 5% of companies are implementing AI correctly in their projects. In other words, 95% are not succeeding.

This is neither new nor surprising. Everyone is hooked on ChatGPT or similar easy-to-use tools that are initially useful; but this remains at a personal, intimate level of use. The real challenge is giving continuity to the implementation of AI tools, making them a meaningful part of our daily routines — and those of our collaborators.

So, what next? First, we must understand the rules of the game. These applications have been trained on information (words, images, even sounds/voices) uploaded to the internet over the years — catalogued and fragmented meticulously to create learning models. Once built, rules (the famous algorithms) were applied to generate new results inspired by those original inputs.

Today, most AI applications (ChatGPT, Midjourney, Runway, Gemini, etc.) compete on usability, offering tools with highly accessible interfaces. The more features, the simpler the use, the faster the responses — and the better positioned they are in the licence sales hierarchy. As for quality of results, that needn’t even be mentioned: as professionals, we can recognise it immediately.

What many don’t realise is that ChatGPT forgets things and hallucinates — and it’s programmed so that we don’t notice. Shouldn’t it be the other way round? This could put us in serious trouble.

What almost no one talks about is the need for personalisation in AI applications: tools and models adapted to our way of working and our style. In fact, the most popular apps now compete precisely in this — in making us believe they adapt to what we need. Let’s be clear: these companies live off meeting the needs of the majority, not of a handful of designers, typographers, or multimedia artists.

This is why we have long advocated for tailored applications and, above all, for the creation of more personalised models. Being able to train your own visual model, or create your own text-generating workflow, should already be standard practice across all sectors. The great advantage in our “world” is that our talent ensures these uses achieve another level of outcome: our level.

For example, we can set up workflows for “intelligent” agents to respond according to specific criteria. Or imagine you are a copywriter wanting to train your writing style. Should I do it with ChatGPT? I wouldn’t. I’d “launch” it across several language models at once (ideally competitors) in a semi-automated way. You still review the results, but the models will have supported a huge part of the production process — with knowledge no one else has (not even daily AI users).

Another interesting case is static imagery, or even video generation. Sometimes it’s handy to create visuals using Freepik or Sora, but in many situations, we need something more artisanal, personal, and aesthetically unique and controlled. In such cases, we have no choice but to turn to more advanced tools like ComfyUI or Forge.
Or, more exciting still, we can finally build our own tools with code (even without being programmers, but knowing what we’re doing). Small interactive elements, add-ons for our websites, or more creative experiments linked to specific branding projects.

We know: we’re short on time, everything feels chaotic, and it’s daunting to start. And yes, we must study, but thankfully the learning curve is much faster than when we had to learn Photoshop from DVDs. Now, in just three months, we can already integrate AI into our processes.

Here’s another striking line from the MIT report mentioned earlier: AI is not reducing jobs. What it is reducing is external collaborations. This means providers must be more efficient and deliver more value than ever to remain in the “positive” 50%.

Of course, because this is our profession, we recommend doing this hand in hand with specialised consultants: professionals qualified in using these tools, who understand their philosophy, their ethical implications, and who also have a 360-degree vision. With proven experience, with an understanding of departments and roles. Professionals who can help us integrate, but also say NO to certain “novelties”.

Together with BCDF we are working on the next residency focused on text-based AIs, where we work with LLMs (large language models) along the lines I’ve been discussing in this article. There is still a lack of focus, and we intend to provide it. Beyond residencies, we can also work on personalised in-house projects. These help launch, in record time, the use of new methodologies and tools; they require understanding both the company’s operations and the concerns of each worker. And let’s not forget that in design, art, and creativity, we must ensure quality in our work is never lost.