What an unspecific buzzword AI is.
We live at a time in which our lives are being transformed by AI. Yet, there seems to be very little clarity on what this technology is, what it’s limitations might be, or how it will affect the world. What is clear, is that a great wave of change is coming and disrupting industries.
In my experience with waves, perspective and preparation are the deciding factors in whether a wave is good or bad. So let’s talk about why you should get your surfboard out and catch this one, rather than fortify your sandcastle!
What AI are we talking about?
Generative pre-trained transformers – the “GPT” that comes after the “Chat-” – are a form of “AI” technology that has rapidly become popular over the last 6 months.
In 2017 Google released a paper that introduced a new type of neural network architecture called “Transformer”. This is the core technology behind most new generative AI. Think of transformer technology working like a skilled chef preparing a recipe.
They take a complex dish (input data), break it down into individual ingredients (smaller segments), determine the importance of each ingredient (assigning importance to each segment), and then use this knowledge to create a delicious and well-balanced meal (generate coherent and contextually relevant output).
A well trained chef can then work with whatever ingredients they’re given, making a recipe that, based on their (pre-)trained guesses, will work.
What are the current limitations and considerations?
Transformer technology can be continuously built upon and improved, but their main limitation is fundamental to how they function. Because they are pre-trained – admittedly on incomprehesible quantities of data – they are limited to only what they have been trained on.
That means that everything that you create will be in a sense, derived from what they have already learnt.
So let’s contextualise that limitation in design. Transformers, by their nature, generate derivative design.
The idea of derivative design is not exclusive to AI. Throughout history, designers have been influenced by and built upon the work of their predecessors. Artistic innovation often emerges by combining and reinterpreting existing ideas and techniques.
However, AI introduces new considerations because it can automatically generate large volumes of work quickly, based on existing examples. This raises questions about the originality and creativity of AI-generated elements in design.
Can AI make good design?
Defining good design is a nuanced matter. While the ability to create visually appealing outcomes may be one aspect, it doesn't solely encompass good design. Even before the emergence of this technological wave, copying and pasting ideas was always possible, but by many, it’s not considered good design – even when it looks incredible.
In many cases, generative AI may outperform humans in creating visually pleasing designs. However, I believe that good design goes deeper and encompasses more than aesthetics.
Value lies not solely in the ability create something, but in the underlying intentions, reasoning, and the evaluation of whether it was a beneficial course of action.
Returning to the chef analogy, when faced with a problem, we need to find the right ingredients and determine the method; that’s what I think good design is and I think there’s room for the use of generative AI in that process.
Final question: What does this mean for me, and the rest of the world?
Generative AI is great at making beautiful, surface level design, but as we’ve discussed, good design goes deeper than aesthetics.
With the emergence of technology that can quickly generate surface level design, we can predict a couple of things: Beautiful but meaningless design will spread further than ever (technology often means ‘more’ not ‘better’), and those who previously offered this as a service may struggle to compete a technology that can offer the same thing quicker and cheaper.
If that worries you, my advice is to design things in a deeper way. Design with intention and purpose beyond aesthetics, and generative AI will go from a scary wave about to crash down on you, to an opportunity for an exciting ride.
*All poetry in this article was assisted by ChatGPT, all poeticism was human inspired.