Interaction Design in the AI Era

Entering the AI Era

I’ve been thinking a lot about what it means to be an interaction designer in the AI era.

Not just how to use the tools. That part feels relatively clear now.

Use them every day. Learn them. Push them. Break them. Prototype with them. Let them accelerate the tedious parts, sharpen early thinking, and help you make more than you could make alone.

But the more interesting question goes beyond productivity.

How do you grow as an interaction designer when the interface itself is being redefined?

That question has been sitting with me because this moment feels both exciting and unstable. AI could make us more capable, more creative, and more prosperous. It could also flatten craft, replace jobs, flood the world with average output, make people dependent on systems they do not understand, or tear apart parts of the work economy before we have a healthy replacement.

Some days it feels like a breakthrough.

Some days it feels like a blur.

And somewhere inside that blur is the real career question: where do I stand out?

Every Era Teaches a New Material

When I look back at the history of software, interaction designers have always had to evolve with the dominant material of the era.

In Web 1.0, the material was mostly information. Pages, links, navigation, publishing, directories, forms. Interaction design was about making information accessible and understandable.

In Web 2.0, the material became participation. Profiles, feeds, comments, likes, networks, sharing, user-generated content, reputation, and community. Designers had to understand not just what people clicked, but why they contributed, returned, trusted, shared, and identified with a product. Tim O’Reilly’s idea of Web 2.0 as “the network as platform” captured that shift well. The product was no longer just something a company published. It became something shaped by the people using it.

Then mobile changed the material again. The interface moved from desks to pockets, from mouse clicks to fingers, from sessions to moments. Software became more intimate. It lived in cameras, maps, sensors, push notifications, location, commutes, restaurants, sidewalks, and beds. The designer’s job became less about pages and more about context, interruption, speed, physicality, and habit.

Now AI is pushing interaction design into another shift.

The new material is intelligence, or at least something that behaves like it. Language, memory, prediction, automation, agents, generated interfaces, probabilistic outputs, and systems that can summarize, decide, act, and create.

That is a very different material.

It is not enough to make the screen look good. It is not enough to understand flows. Interaction designers now have to understand intent, trust, control, uncertainty, delegation, reversibility, and human judgment.

The interface is no longer just where a person clicks. It is where a person negotiates with a system that might act on their behalf.

I’ve Lived Through Smaller Versions of This Shift Before

Part of why this moment feels familiar is that I have felt versions of it before.

In the Web 2.0 era, the barrier to making things felt low enough that an interaction designer could jump in. You could pick up PHP, hack together a bulletin board, set up a community, customize a template, wire up a simple database, and suddenly you were not just designing the surface. You were shaping how the thing behaved.

It was not always elegant, and it was not always scalable, but it was accessible. That mattered.

It gave designers a way to understand the medium from the inside. You could see how a form became a post, how a post became a thread, how a thread became a community, and how small interaction choices changed participation. You could learn by making. You could break things, fix them, and understand the relationship between interface, behavior, and people.

Then mobile came, and I remember leaning into that too.

At InfoScout, I learned Xcode and embraced the mobile era as something I could grow into technically. It was surprisingly easy to pick up at first. Interface Builder, screens, flows, view controllers, transitions, a lot of it mapped naturally to how interaction designers already thought. I could prototype. I could understand the constraints. I could speak more fluently with engineers. I could make ideas feel more real.

But I also hit a wall.

Backend work was different. Once I got deeper into infrastructure, data models, APIs, scaling, and the invisible machinery behind the product, the learning curve got steeper. I could understand pieces of it, but it did not feel as naturally close to interaction design as the front-end and mobile layer did.

That wall taught me something important: each era has an entry point for designers.

For Web 2.0, it was simple frameworks, PHP, forums, templates, and participation systems. For mobile, it was Xcode, touch interfaces, and prototyping native experiences. For AI, the entry point is broader. It is prompts, prototypes, agents, APIs, code generation, model behavior, workflow design, automation, evaluation, and new interface patterns that have not fully settled yet.

That is intimidating.

But it also means the door is wide open.

AI Makes the Field Broader, Not Narrower

One fear I have is that AI will make design feel cheaper.

If anyone can generate a screen, write a flow, create a logo, produce copy, or mock up a concept, then what happens to the designer?

But maybe that is the wrong question.

Maybe the real question is: what becomes more valuable when average output becomes free?

AI is very good at generating the middle. The expected answer. The average layout. The obvious headline. The familiar pattern. The safe concept. The thing that feels like everything it has already seen.

So the opportunity is not to become average faster. The opportunity is to become hard to average.

That means building taste, judgment, technical curiosity, human understanding, and a point of view. It means knowing how to frame the right problem. It means seeing what is missing. It means noticing when the AI is technically correct but emotionally wrong. It means designing systems that help people feel more capable instead of more dependent.

That is where interaction designers can stand out.

Not by becoming prompt operators. Not by using AI as a shortcut around thinking. But by becoming designers of agency.

The New Interaction Design Questions

In the AI era, the best designers will ask different questions.

Not just where this button goes, but what the AI should actually be allowed to do. Not just whether the experience feels fast, but where it should slow down and ask for confirmation. Not just when the system should speak up, but when it should stay quiet and let the user stay in control.

That leads to a different set of concerns: what happens when the system is wrong, how people recover when it misses the mark, what it should remember, what it should forget, and how much of its thinking should be visible along the way.

These are not edge cases. They are the core interaction problems of intelligent systems.

A good AI product is not just powerful. It is legible. A good AI product does not just complete tasks. It preserves agency. A good AI product does not just feel magical. It knows when magic becomes dangerous.

That is why interaction design matters so much right now. As software becomes more capable, the experience layer becomes more consequential. The interface is where power becomes understandable, usable, safe, delightful, or harmful.

Augmentation Is Still the Better Frame

One of the most useful ways to think about AI is through the older idea of augmentation.

Doug Engelbart was not trying to replace human intelligence. He was trying to augment it, to increase what people could understand and do. That distinction still matters.

A lot of AI products today are framed around replacement. Replace the writer. Replace the designer. Replace the analyst. Replace the researcher. Replace the assistant. Replace the team.

But the more interesting design space may be augmentation.

How do we help a designer see more options without losing their taste? How do we help a researcher find patterns without flattening nuance? How do we help a product team explore tradeoffs faster without pretending the tradeoffs disappear? How do we help a person make a better decision, not just a faster one?

In other words, how do we make the human stronger?

That is a different posture. And it is one interaction designers should fight for.

The Designer’s Role Is Moving Closer to the System

For a long time, designers could operate mostly at the interface layer. We could make wireframes, flows, prototypes, design systems, journeys, and handoff specs.

That work still matters. But AI pulls us closer to the system underneath.

The behavior of the product is no longer fully deterministic. The same input can produce different outputs. The quality of the experience depends on context, model behavior, memory, retrieval, permissions, fallback states, and confidence. That means designers need to get closer to how the thing works.

We do not all need to become machine learning engineers. But we do need to become more technically fluent. We need to understand what models are good at, where they fail, what hallucination means, how context changes output, how agents plan, how tools get called, how data flows through a system, and how to evaluate whether an AI interaction is actually useful.

This feels similar to learning Xcode in the mobile era, but bigger.

Back then, learning the tool helped me understand the medium. It helped me design better because I understood the constraints. Now, learning AI tools, APIs, and agent workflows can do the same thing.

The point is not to become an engineer. The point is to become a better interaction designer because you understand the new material.

The Twist: AI Makes Originality More Important, Not Less

The more I think about it, the twist of this era is that AI does not remove the need for originality. It raises the value of it.

When everyone has access to the same tools, the tool itself is no longer the advantage. The advantage becomes what you bring into it: your taste, your references, your instincts, your lived experience, your ability to notice what feels off, and your willingness to keep pushing after the first decent answer.

That is where the playing field feels wide open.

AI can give you more directions than you could explore alone. It can help you prototype something you would have only described in a meeting. It can help you learn enough code to cross the gap between idea and working thing. It can help you study references outside your usual lane, pressure-test assumptions, and find the obvious answer faster.

But finding the obvious answer faster is not the same thing as doing original work.

In fact, the danger is stopping there.

AI is very good at getting you to something that looks complete. It can make an idea feel polished before it is actually interesting. It can smooth over the rough edges that might have led somewhere stranger. It can make everything sound confident, even when the thinking underneath is thin.

So the craft becomes knowing when the answer is only average.

The craft is knowing when something is too smooth, too expected, too familiar, too much like everything else. The craft is knowing when to reject the first nine options because the tenth has a pulse. The craft is using the machine to move faster through the obvious so you can spend more time on what is specific, strange, useful, beautiful, and true.

That is where originality lives.

Not in pretending AI does not exist. Not in using it as a shortcut to skip thinking. But in using it as leverage to think further.

Closing Reflection

I do not want to stand outside this change and comment on it from a distance. I also do not want to blindly chase hype.

I want to be close enough to shape it.

That means using AI daily, but not lazily. It means building small things. It means prototyping ideas instead of just talking about them. It means learning enough code and systems thinking to collaborate better. It means writing about what I am seeing. It means studying the history of interaction design so I can recognize patterns when they repeat.

It also means staying human-centered in a moment when the technology can make people forget the human.

The designers who grew in Web 2.0 understood participation. The designers who grew in mobile understood touch and context. The designers who grow in AI will understand agency, trust, collaboration, and originality in a world where generation is cheap.

That is where I want to grow.

Not just as someone who uses AI, but as someone who can use it to push the craft forward.

The AI era is not asking interaction designers to disappear. It is asking us to evolve again.

The craft is changing, but the deeper responsibility is the same: make technology understandable, useful, humane, and empowering.

Maybe the screens change. Maybe the tools change. Maybe the interface becomes more conversational, ambient, agentic, generated, or invisible. But people will still need to understand what is happening. They will still need control. They will still need confidence. They will still need tools that respect their judgment. They will still need experiences that make them feel capable.

That is the work.

And if average output becomes infinite, then average is no longer enough. The value moves to the person who can direct the machine toward something more specific, more useful, more beautiful, more humane, and more alive.

The future of interaction design will not belong only to the people who use AI the most. It will belong to the people who use it with the most intention. The people who can bring together craft, systems, behavior, technology, and feeling. The people who can design not just what AI produces, but how humans grow through the experience of using it.

The job is not to keep up with the machine.

The job is to use the machine to push the craft further than before.

And in a world where everyone can generate, the real challenge is to become more original, more intentional, and harder to average.

Best,

Craig Aucutt