Designing Responsibly in the Age of AI: What We’re Learning as We Go
Who needs another blog article about AI and the new way of working? Nobody, probably. The internet is already oversaturated with opinion pieces, hot takes and voices that have it all figured out, meanwhile technology and the way we work with it changes constantly which makes it really hard to navigate all the information online.
That’s why I’m not writing this about technology but about us as humans and how we deal with it. In my experience the perfect antidote for all the information overload is talking to actual people working through the same questions.
A great opportunity to do this was the Design Leadership Meetup hosted by Boye & Co in Hamburg a few weeks ago.
AI has changed the industry more than maybe anything ever before and will continue to do so. We don’t have all the answers and won’t pretend so but here are a few things I took from our Meetup.
Who’s Actually Responsible Here?
AI can produce any output from design to code but who actually feels confident to sign their name on these outputs? In a time where anyone can produce anything with the help of AI, who ensures quality and security? Without clear ownership, quality and security erode—not dramatically, just quietly.
I’m a Designer with a fairly good understanding of technology, for me code feels similar to Dutch: I can read most of it but I can’t speak it. So I can read code but I couldn’t write it and I also wouldn’t feel comfortable pushing something AI wrote to a live environment.
As a Developer or Project Manager it’s similar: Even with tools like design.md and a good personal understanding of design, who catches it when AI loses brand intent or UX nuance? Who would feel responsible for it?
For me this means we still keep our separate roles with our own expertise for now, so someone can feel responsible during each phase of the project. However we can’t be sure that this won’t change in a couple of months so we will keep extending our technical knowledge as designers and think even more about aligning design and code.
Design Documentation Has to Be Airtight
AI amplifies what’s already broken. Things we could balance out as humans become more visible when using AI. Gaps in Design Documentation are just one example for this. Vague guidelines and inconsistencies build up like a snowball, from one small oversight to unreliable results that need loads of refactoring. These aren’t new problems, we were just able to work around them as a human team. AI makes them impossible to ignore.
Design Documentation has already become more and more important than over the last couple of years as our systems have become more complex. We at b13 also work with our clients over long periods of time and our projects are ever evolving so it has always been important to keep our designs well documented and consistent but now our documentation needs to be even more precise, structured and aligned to code.
As a side benefit, precise design docs can also be used to test implementation over time, making sure we don’t lose track of our original designs.
Using AI Responsibly (Without Pretending We Have All the Answers)
Environmental and social costs of AI are real and widely known—but avoiding it entirely is increasingly unrealistic. It can be really easy to get swept away by excitement about all the possibilities AI presents but the risks and negative impacts should never be ignored.
We also become increasingly dependent on a few big tech companies who can and will change pricing, terms and availability.
In addition, the human side is just as important: job anxiety and AI burnout is real and people need space to process it and voice their thoughts and fears.
We believe that talking openly is better than letting everyone figure it out alone. We want to create an atmosphere where it’s okay to not know something, okay to ask questions and feel insecure. In an ever changing and fast paced world it has become a luxury to be able to say “I need to sleep on it” but we want people to be able to do so.
All this being said, while it’s important to know all the risks and consequences of AI, we also want everyone to feel on board and not left behind. To balance exactly this, our host in Hamburg has developed a great method: the Responsible AI Canvas. It provides a structured way for teams to weigh the pros and cons together, giving everyone a voice and a place to speak up. I highly recommend checking it out at https://truepriceofai.org/ai/responsible-ai-canvas.
Where We Land (For Now)
We don’t have it all figured out and I won’t give you another opinion piece pretending to know the new best way to work in the Age of AI. But here’s what we do have: a commitment to empathy and the humans in our team and a curious and open mind to keep asking (hard) questions. Our goal is not to replace people with AI, but rather to automate tedious tasks that no one wanted to do in the first place. To achieve this, we will continue to meet in teams, discuss our workflows, and identify bottlenecks and minor annoyances that can be eliminated—or at least mitigated—with the help of AI tools. This doesn’t mean We’ll mindlessly turn to AI for every problem—we will carefully consider exactly what we need. Is it a simple chatbot, a skill, or an agent? We’ll always keep the risks and price in mind and try to strike the right balance. It’s certainly a process and a challenge, but we believe that as a team, we can figure it out.
We will keep talking, keep listening and iterate as we go.
You have hard questions relating to AI?
We found an outside perspective like this Meetup helped us to gain insights on how to go forward. We’d be happy to schedule a call discussing your requirements and objectives to get AI-ready.