Been on social media lately? Scroll through your platform of choice and you’ll find headlines like these:
‘The design process is dead’. ‘The one-day-website workflow’. ‘AI has completely cooked UX designers’
Nearly half of UX professionals feel insecure about their future – a figure that, according to a 2025 industry survey, jumped 26 points in a single year. That’s not a statistic. That’s fear, written with numbers.
And on the face of it, fear seems like a valid reaction. We’re faced with something faster, cheaper, and relentlessly more capable. Something that doesn’t take lunch breaks or wellness days.
After all, did horses ever stand a chance against cars?
But perhaps fear needs to give way to reflection. Because the real concern isn’t hiding on Instagram or LinkedIn. It’s hiding in plain sight.
When design becomes automated, we don’t risk losing jobs. We risk losing judgment, craft, and the friction that produces insight. We risk losing the imperfections, the rough edges, and the quiet genius that makes us human – in service to other humans.
Remember, horses couldn’t make cars better.
Chapter 1: The Quiet Erosion
The Illusion of Competence
Humans have always deferred to systems.
Aviation, healthcare, defence – decades of research show that when an automated system makes a suggestion, we follow it. Even when we shouldn’t. Psychologists call it automation bias. The rest of us call it trusting the autopilot so we can put our feet up and play Candy Crush.
But here’s what’s different about AI. Traditional automation operated within guardrails we designed and understood. AI doesn’t just work within those guardrails – it writes new ones. Quietly. Continuously. Often without us noticing.
And it’s remarkably easy to trust. Because AI-generated outputs look confident. Designs look polished. Copy sounds considered. Workflows feel logical. Everything looks right – and that’s precisely where the risk lives.
Polish creates perceived authority. But polish isn’t understanding.
Would you trust the autopilot when the plane loses an engine?
Research decay
I don’t usually look to horror movies for life lessons, but they did teach me one thing – shortcuts often lead to the wrong end of a chainsaw. Or a hatchet, depending on your franchise of choice.
AI can generate personas, suggest user flows, predict likely behaviour, and even attempt something it considers emotionally-driven design. And as teams start using it to do exactly that, they also start skipping primary research because the model has done a passable impression of it.
Before long, we’re building solutions on maybes and most-likelys.
The problem isn’t the tool. It’s the assumption underneath it – that synthesised behaviour is close enough to lived experience. That a model trained on patterns can replace a conversation with a real person having a genuinely bad day.
It can’t. Try explaining the mind-numbing trauma of a 4-hour long IT training workshop to a chatbot. Empathy requires exposure, not inference. And the moment we automate it, we’re not designing for humans anymore. We’re designing for a cheap approximation of them.
Let’s leave automated empathy for sci-fi novels. It fits in better with killer androids and sentient spaceship AIs anyway.
The Emotional impact of automation
Here’s a dynamic nobody talks about enough.
When designers feel replaceable, they compromise. They move faster. They ship more to prove they’re still relevant. They stop pushing back on briefs and they stop asking uncomfortable questions. They optimise for output because output is visible, and the rigour that made it possible isn’t.
And in doing so, they accelerate the very erosion they’re afraid of.
It’s the professional equivalent of stress-eating. The thing you do to feel better knowing it’ll only make things worse.
We were never competing on speed. A junior designer with a half-decent prompt can produce twenty layout options before lunch. That race was lost before it started.
What can’t be replicated is the conviction to throw all twenty away because none of them are asking the right question. That judgment – the self-belief to start again – is exactly what’s worth protecting.
If we define our value by velocity, we’ve already lost.
Chapter 2: The Deliberate Hold
The pause
Ask two designers the same question, and say they give you the same answer – identical in every way. Except one took their time, sat with the problem, and came back to you. The other responded immediately.
Who do you trust more?
There is quiet power in the pause. It signifies reflection, not hesitation. The kind isn’t tracked on a timesheet, but shows up everywhere in the quality of the outcome.
When we rush straight to polished outputs, we skip the messy middle. And the messy middle is where understanding actually forms – where the obvious solution is discarded and something better emerges from the discomfort of not knowing, and the strength to accept that realisation and relinquish control.
AI doesn’t understand that middle. It moves from prompt to output in seconds, skipping the friction entirely. That might be useful when the friction is mechanical, but it’s a problem when the friction is of a human nature.
There’s a reason the best ideas hit you when you least expect them to. They arrive in the shower, on a walk, at 3am when you weren’t trying. Slowness isn’t inefficiency. It’s incubation.
Interrogate. Question. Redraw. The insight is in the resistance.
Craft through human error
Craft isn’t making things pretty. It’s the thinking, the doing, the wrong turns, and the questions that follow when something doesn’t work.
When you rewrite a piece of copy for the fourth time, you’re not failing to get it right. You’re getting closer to understanding what right actually means. That’s not inefficiency. That’s rigour.
AI reduces repetition, mechanical errors, and production time. That’s genuinely useful. But it also tends toward the statistical norm – the output that is neither shocking, nor surprising. It’s like going to your favourite gelateria and discovering every flavour is vanilla. But hey, at least they’ve got five different toppings to keep it exciting.
Human error doesn’t work like that. We are naturally wired to find beauty in imperfection. Kintsugi – the Japanese art of repairing broken pottery with gold – doesn’t hide the cracks. It celebrates them. The cracks became the highlight. The mistake became the product. The deviation became the innovation.
Design was always a mirror into humanity – original, different, imperfect, and better for it. Let’s not discard the very imperfections that make design meaningful – but try and understand them.
Because the wrong answers have always led us to the right questions.
The weight of morality
CEOs protect profit. Marketers chase leads. Influencers count likes. Every role has a singular metric that defines success – and their relentless pursuit of it.
But strip away the titles and the KPIs – what are we left with?
Perhaps the most important thing we do – regardless of role – is help empower humans to make informed decisions.
Able humans. Vulnerable humans. Secure humans. Unsure humans.
We live in a world where filming incidents on our phones feels more natural than stepping in and taking care of the situation. Psychologists call it the bystander effect – the more people present, the less any individual feels responsible for acting. Diffusion of responsibility. Everyone assumes someone else will step in.
Automation works the same way. The more layers between a decision and its human consequence, the easier it becomes to feel like the outcome isn’t really yours.
If AI suggests a dark pattern, and a product manager approves it, and a designer builds it, and an engineer ships it – who owns the moral weight?
The more we automate the process, the harder that answer becomes to hold onto.
Accountability doesn’t diffuse just because the pipeline does.
The wrap-up
I was terrified the first time I rode a bike. And then on came the training wheels. They didn’t just stop me falling – they gave me the confidence to move forward until I didn’t need them anymore. Temporary guardrails. Not a permanent crutch.
That’s what intentional design practice looks like in an automated world. Not a rejection of the tools, but a decision about what we hold onto while we use them.
Research over assumption – because understanding a person is more valuable than predicting what works for them.
Craft over convenience – because the thinking that happens in the making is irreplaceable.
Reflection over velocity – because the fastest answer and the right answer are rarely the same thing.
These are the guardrails we set for ourselves – the things we protect not because automation can’t touch them, but because we’ve decided they’re worth protecting.
Solving human problems has always been a human responsibility. And the more we examine what automation does to that responsibility, the more we understand something larger about ourselves.
Maybe that’s where this ends up. Not just in the humanity of design – but what design reveals about the design of humanity.
Because no one is ever too old for training wheels.