AI is yet to find a foothold in modern design, as algorithms can’t account for human messiness

Nitor’s Senior Designers Paula Ikonen and Annika Madejska are both seasoned AI users. For Ikonen, however, most deep dives into the world of generative AI have been taken during off hours:

“In terms of work, AI has represented a relatively small change up to now. The project I’m currently working on doesn’t require AI implementation. But in private, I’m quite addicted to it. Midjourney has been a part of my life for over two years, and I’ve experimented with AI in everything from videos to music.”

For Madejska, AI has already found its way into her daily work:

“I’ve had the rare opportunity to apply my designer skills specifically to AI product design. In this case, it was very specialised software for experts to train their own AI models within the health sector. I operated as the bridge between medical experts, data scientists, and developers. The goal was to ensure the final user interface would be as usable and intuitive as possible without compromising on the product’s expert level usage and high bar of entry.”

Current AI isn’t designed for collaboration

Bridging the proverbial gaps between concept, production, release, and use case analysis comprise an inherent part of a designer’s workflow. Success requires not only a keen analytic eye and capacity to distil vast amounts of information into usability and practicality, but also social skills, communication, and cultivating collaboration.

Dealing with the human messiness of it all, as Madejska wittily puts it. She continues:

“As designers, we need to be able to engross ourselves in the fundamentals. What can we do with the data? How does the data need to be organised, structured and categorised to make sense? In terms of AI, that’s also a core question we find ourselves asking. Why are we adding AI to this, why is this model chosen? What problem does it solve?”

A key component of design work is understanding the challenges and ambitions of clients, and what that entails in terms of operations. Often the work has more to do with behavioural science than drawing UIs. The root causes of problems need to be uncovered before they can be solved, and business goals cannot be met without laying proper groundwork before proceeding. This entails constant dialogue and teamwork – human wants, human needs, human ideas. Mediating, communicating, finding out.

For designers, a key issue with most current generative AI interfaces is that they are designed for single users. This can make cohesive teamwork difficult, hindering both problem-solving and innovation. AI platforms are slowly adopting new UI possibilities to enable more fluent co-working, but the current state is still far from ideal. Another key issue is the sheer amount of AI tools available – the question is how to uncover the needles in the mountainous haystack that will prove truly useful for a designer in the midst of solving a UX problem or accessibility issues.

An ethically troublesome web of algorithms

Our interviewees concur that a general misconception regarding AI is that it can be viewed as a monolithic, all-encompassing structure. In fact, the current usability of AI is still basically machine learning systems working together as socio-technical systems collecting data and performing tasks. When those systems can be brought together in a collaborative chain, they can complete quite complex tasks. However, the effectiveness of the implementation is still dependent on how well individual systems and algorithms can be steered to work together in unison – the realm where human expertise shines.

Madejska also notes the acceleration of discussion around the ethics of AI:

“People in our industry need to be aware of all the upcoming EU regulations concerning AI and technology, and what impact they will have on organisations and professionals. Even if your work doesn’t involve high-risk AI, strategies and governance models need a fresh perspective. Companies operating in the EU may find themselves blindsided by how many new and amended regulations will impact the tech field in the near future, directives like the AI Act being the tip of the iceberg. It is crucial for businesses to understand how these regulations will come into play.”

With recent product liability directives being updated, software will be regarded in the same sense as tangible products like toys and light bulbs. That means businesses and organisations need to be more cohesive in their risk assessment regarding technology and its utilisation.

In terms of design, the question of ethics can’t be dislodged from the AI conversation. Ikonen continues:

“These big, sometimes clandestine black boxes of algorithms can often produce unforeseeable outcomes. Unpredictability can indeed foster innovation, but it makes a designer’s work more complex. We need to be able to forecast possible harmful outcomes to mitigate them before products are released. With current AI systems, it can be challenging to recognise and curate emergent patterns and see their potential. In order to innovate, we as designers have to contextualise them and figure out where to put them to good use. This necessitates careful implementation with human oversight, with monitoring measures included with deployment.”

The great techland gold rush

Ikonen continues by noting that the current AI boom resembles the gold rush in the USA in the 19th century. People rode out to grab any piece of land they could to make their fortune, and everyone involved was in a frenzy to move faster than the others. With AI, the tech industry finds itself in a similar situation.

“We often need to facilitate the curation of massive amounts of text and images. It is vital to be able to step back, slow down, and take the appropriate amount of time to evaluate what has been produced,” Madejska continues.

AI has introduced pressure to produce results faster. Prototyping can be done more swiftly, which in turn means projects can get underway faster. The downside is that speed can overshadow quality. Companies may find themselves focusing too heavily on quick turnarounds and the face value of saving costs, but these benefits often come at the cost of curation and, by extension, the quality of the end products.

Finding a balance between efficiency and quality is rooted in the company’s maturity level, how well all applicable data is handled, and how well it can be utilised. In this equation, AI can speed up processes, but it can also sway companies to focus too heavily on cost-cutting and not enough on the long-term benefits of creating better products and cultivating a lasting work environment.

“I think we have a massive amount of work ahead of us. Governance and organisational structures need to be overhauled, and methods of working need renewal. But beyond that, there is much work to be done in how we collaborate within organisations and between teams, and how we onboard people into utilising and harnessing AI. That is the path towards true innovation,” Madejska concludes.

Interested in turning your company’s data into practical innovations? Read more about our services!