AI has taken the world by storm. While it is not an entirely new concept, it has gained tremendous prominence in recent years. A range of tools like ChatGPT and Midjourney made their debut to the general public, sparking both wonder and concern. The applications are endless, the expectations sky-high. 

But what exactly is AI? What can we expect from it? And how can you start using it today? I spoke with Zack Kinnaer who recently dove into the subject.

‍

Jef: Can you explain what you did at KAN and why you immersed yourself in AI?

Zack: I study product development at the University of Antwerp. After my bachelor's, I joined KAN as an intern last year, which was a very enjoyable experience. This academic year I still had time left in my curriculum and was able to work here as a part-time student. During my time at KAN, I immersed myself in some recent technological developments, such as blockchain, NFTs, the metaverse, and ultimately AI. I explored how we could use these things to serve clients better and enrich our own services. 

‍

Jef: Since last year, AI has been 'the talk of the town', but what is it really?

Zack: It's essentially smart technology that can be used to perform tasks that would normally require human intelligence. It may be a hot topic today, but AI is not really new. We already use it frequently in our daily lives. Just think of translation tools like Google Translate, self-driving cars or robot hoovers, etc. These all use AI. In recent years, evolutions in hardware, the availability of huge amounts of data and better machine learning algorithms have accelerated AI. As a result, today we have remarkable tools like ChatGPT and Midjourney, which have caused quite a stir.

‍

Jef: What do you think is the potential of AI?

Zack: The long-term potential seems to me to be a form of artificial intelligence that is integrated into our daily lives in such a way that we interact with it naturally, without it feeling like technology. But of course, we are not there yet. Today, you can use AI to tackle certain time-consuming tasks more efficiently. For example, if you want to make a report of a meeting. Instead of writing out and summarising notes or a recording file yourself, you can use a tool to process all that information. This allows you to simply review and spend your time on more important matters. Or you can use it for inspiration. If you are stuck creatively for a moment or you are still starting with a blank canvas and want to check some ideas quickly, you can use Midjourney instead of scrolling all over Pinterest, for example.

‍

Jef: I understand you've also been working on a tool at KAN that uses AI to quickly create a representation of a brand. Can you tell us something about that?

Zack: Yes, that's correct. KAN uses its own tool, the BrandScan, which captures a brand's personality through an online questionnaire. This results in a brand archetype, several characteristics, and an indication of the visual style that serve as the foundation for a creative brief. You can actually think of it as a brand's passport. My focus was on translating the results of this BrandScan into mood boards through AI, which bring the brand's DNA to life. It gives a first impression of what the brand could look like in different applications. A designer can then select, fine-tune or adjust the generated concepts through a user-friendly dashboard. By going very broad in terms of ideas at the beginning, several options can be explored in a short time, which speeds up the creative process. This ensures that the client always receives the highest quality.

‍

Jef: Should we be worried that AI will replace human creativity - and thus the job of creatives?

Zack: I personally don't think so. The thing is, designers should never let it get to the point where they let AI completely take over their creativity. It does not deliver flawless designs and should only serve as inspiration. In my opinion, creatives should not see AI as a competitor, but rather as their ally. It enables you to be more efficient, which allows you to deliver more and better work. To return to the question, I don't think AI will take over the job of creatives, but I do think that those who can't work with AI will be overtaken by those who can.

‍

Jef: How can someone start using AI themselves? Do you need certain skills for that?

Zack: A good way to start using AI is to ask yourself whether there are repetitive tasks in your daily activities or tasks that require significant data processing. Once you have identified these, you can look for a suitable tool that can help with this. For example, you could ask ChatGPT, "What AI tools help with [process X]?", and then start working on this. 

This is also the initial hurdle to working with generative AI: prompting. A prompt is an input – often text – that you enter into the AI model to generate a desired output. You can actually compare it to a design briefing: with it, you indicate the direction you want the result to go. The more specific and clear your prompt is, the more targeted the output of the AI model will be.

‍

Jef: So do you have any tips on how to write a good prompt?

Zack: I myself have experimented quite a bit with writing prompts for ChatGPT, and I have learned a few lessons from this. I often ask the model to respond from a certain role, such as that of a top chef if I have a cooking-related question, for example. I also provide an idea of what I expect as a response, like creating a recipe using the ingredients I have in my fridge. Additionally, it can be helpful to give ChatGPT an example to draw inspiration from. For instance, you could ask it to respond like Gordon Ramsay, Jamie Oliver, or Pascale Naessens, depending on your preferences. At the end, I describe what I expect as output. In the case of a recipe, I want a title, a paragraph about the dish itself, a list of ingredients and easy-to-follow steps I can get started with. Then, if the output does not fully meet my expectations – say the measurements are in imperial units instead of metric – I give feedback and ask to rewrite the answer with it.

‍

‍

Jef: The way you say this now, it seems to me that it's more like writing out a task for a student or an intern who then carries it out in seconds, right?

Zack: Yes, you can indeed compare it to that, although we are talking about an intern with an immensely broad knowledge who can answer incredibly quickly. This 'imaginary intern' is just not very good at understanding context or nuances in questions. Nor can you expect initiative or suggestions without giving explicit instructions. I think this comparison also somewhat captures what it means to work with AI. Basically, anyone who wants to can have an assistant with strong analytical capabilities who processes data in a way that mimics human intelligence and creativity. It can help with repetitive tasks or the processing of huge amounts of data, and patiently waits for a command to get to work. You then act as the manager of that AI, stringing together and editing the broad output into a cohesive narrative.

‍

Jef: Curious to see what the future holds! Thanks for these insights, Zack. As Always: Inspiring.

Zack: You're welcome, Jef, and thanks for this cheap pun.