Whether we admit it, like it, or believe it, we’re in a relationship with AI.

It’s in the apps and tools we use daily. It’s in the conversations we have with our friends and family. It’s on the news, affecting our markets, and sometimes hiding in plain sight.

Some of you might be embracing this new relationship excited by the shiny new efficiencies it brings to your life.

Others might be in denial about this relationship, unwilling to admit its effect on you. We all know where that story ends.

And for many, you might be fearful about this relationship and the change it represents. The feelings of 

 it might surface. Those thoughts and fears are not yours alone. There’s been a growing undertone of fear and chatter in our industry that AI is coming for our jobs.

I understand the fear, but refuse to be petrified by it. Not because the threat isn’t real but because it isn’t productive.

There are some jobs that only humans can execute that AI will not replace. That’s because AI cannot replace the thing that makes us human. Emotions. Empathy. Intuition. Creativity.

It can be programmed to fake it and is getting better each day. This means designers must continue to step up our game, use the latest tools, and learn new skills. We shouldn’t focus on skills that will eventually be automated. Using the right tool for the job will be more important than ever. That includes designing 

Steve Jobs said computers are like a bicycle for the mind, and today’s pace feels faster than any bike. I agree with my friend and client Ravi Mehta when he says AI is like a motorcycle for the mind.

We need to be mindful of our direction to avoid losing control or falling. And we should create with principles that stand up for meaning, depth, transparency, and relevance.

I’m an optimist, admittedly, and believe most people who work in tech got into this business because we care about people, and want to make the human experience better.

We must develop and nurture our relationship with AI so the products we create don’t lose that human touch — or worse cause human harm.

Attention has turned into a scarce commodity as massive corporations and early-stage startups alike fight to capture ever more of it. Focus and time are the new luxury.

The competition is about to get fiercer as AI slashes the costs of creating the materials needed to keep us paying attention — digital content, graphics, and products can now be produced with a fraction of the time and effort previously required.

Layer in generative interfaces, and gone will be the idea of the same mass-produced interface for everyone. Instead, interfaces will understand the unique context of the user, their preferences, reflecting back who they are and what they consume. Only those who create compelling experiences will win, triggering five fundamental mindset shifts.

Shift #1: Content Overwhelm → Unique Curation

Generative AI tools can 100x output, meaning we’ll see an even bigger flood of boring, superficial, copycat content. It will be even harder for brands to connect with audiences through this noise.

Sick of mass-produced everything, we’ll seek out the bespoke — and accept nothing less. Makers who curate unique experiences tailored to users’ specific wants, desires, and needs will win.

Take Replit, for example. Offering personalized coding environments with AI-powered tools with community-driven content, they empower developers to code smarter and faster with tailored suggestions and real-time teamwork.

David Hoang, former VP of Design at Replit, is now taking this idea to scale with Atlassian. We’ll start to see more design leaders emerge who harness the power of AI and bring its inherent value to businesses, while maintaining a human touch to everything they do.

I’m watching this space as companies like Atlassian launch Rovo, an AI tool to unlock organizational knowledge. Alongside Agents to increase developer productivity, like Autodev and Autoreview, the idea of AI “doing our laundry” to make room for our creative pursuits is clear. You can see their plans clearly: to support the human in the interaction.

Shift #2: Attention Economy → Meaning Economy

The abundance of clickbait and one-size-fits-all products will drive up the premium on meaning. We’ll expect deeper, more meaningful experiences than ever. The bar will rise massively for every brand and storyteller.

. “Move fast and break things” is too risky for today.

We’ll need to create experiences that provide real value and significance. Experiences that give users tangible outcomes and meaning. Recently, I was talking with my friend Cameron Koczon, and we agree that designers will need to design for depth, not the scale of years past.

I’m not worried about this challenge, because we do this naturally: the majority of designers are deeply sensitive and empathic.

Many of us distrust how AI companies use our data (we can thank 

As long as you were mildly paying attention to what’s been happening with design tooling this year, you may have seen 

 started by Andy Allen of Not Boring. The harsh reality is, as a user of any piece of software, you run the risk of your data training the machine. Open AI bluntly tells users “Don't share sensitive info. Chats may be reviewed and used to train our models.”

At least they’re being transparent with this one.

Deep reads of EULA and simple “turn training off” toggles won’t suffice. The problem is users are locked into certain ecosystems and will continue to be until there’s enough fragmentation and democratization of tools to choose how you build/what you use.

Those who empower users to both understand and control how their data is used will stand out in crowded markets.

Speaking of transparency, recently a conversation on our internal Slack talked about the possibility of the blockchain verifying the source of videos, combatting deepfakes. This would involve technology that’s already available, and the subscription for the on-chain video software you’d be paying for anyway can cover the costs of this process.

Simple solutions can also work. Official domains become more important again. So will in-person meetings. Trusted curators and intermediaries will verify materials with their reputations. We’ll go beyond paying for the checkmark and instead demand authenticity with our purchasing power.

. And that should be people’s prerogative. This concept isn’t new, 

. Companies who respect user’s boundaries will win in the long run.

The novelty of AI sparkles will soon seem antique. AI fatigue will fade as it becomes integrated into everything. No longer a differentiator, AI will be the norm.

AI-first products and companies will need something other than “we use AI” to differentiate them. Gimmicky products that are not much more than wrappers built around ChatGPT or Claude will disappear. Unless you deliver a 

 user experience, big AI companies will eat your use case for breakfast.

This isn’t to discourage makers from building from known models to learn and iterate; the point is that applying world-class UX to the killer use case will be the differentiator, and ultimately the winner.

Shift #5 Pixel Perfect → Generative Systems

With large language models in the mix, UI kits and Content Mapping become more important than rigid page designs or templates.

When we were designing Suno’s mobile app, it quickly became clear that we were never going to be able to design for every edge case. So we shifted our focus to creating and stress testing a robust set of component libraries.

We still delivered hi-fi designs for key features and flows but with the understanding that they would quickly evolve and be used to demonstrate a successful execution of the system. Systems will generate as the content evolves bringing a fluidity back to design we haven’t seen since the earlier days of the internet. Living and *almost* breathing as they generate.

How to design for a better relationship with AI

While AI has been assisting us for years, recent advances in generative AI are exponentially increasing the possibilities of what we, as product designers, can achieve. But the fundamentals are still the same as ever: a product needs to solve a user problem in a delightful, intuitive way.

Instead of giving AI the work we enjoy, let’s use it to help us solve the problems we struggle with (or dread). Let’s use AI to bring humanity to opaque, technical areas like business planning and data analysis. Or doing your laundry and dishes. (Even though I’m not sure I’ll have a robot in my house anytime soon.)

 — it lives within all the apps you use every day (think: Slack, Zoom, Google Docs, LinkedIn), acting like a personal OS. Augment compensates for those things most of us struggle with: organization and memory recall between and within all those apps.

This design problem isn’t new, but AI powering the organization of “just in time” information will be a productivity game changer. 

Perplexity uses a human touch to curate the best sources, and cites every response for answers users can trust. They also don’t let the conversation stop by introducing relevant related questions to paint a fuller picture. I have a lot of respect for people who learn from users and build products that respond directly to their needs (or gripes) when it makes sense.

Taking a look at just a couple of these best-in-class examples, we can clearly see when we let AI do the busy work, we make space for human creativity. No matter what you think of GenAI tools designed for creative output, like Midjourney, Pika, Sora, or the like, you can’t reasonably argue against the need for making more space for human creativity.

Don’t compete on tech sophistication; compete on user experience. Make your value prop clear to users by clearly answering “what’s in it for me?” Then fulfill that promise in the product by helping users achieve tangible outcomes.

 is on a mission to change how we age. They aim to help users slow aging at the cellular level and increase energy.

And that’s what they do. 62% of Tally Health members have been able to reduce their epigenetic age by an average of 2.34 years, while 75% of members have increased their lifestyle survey scores, a measure of how well they are doing across five key wellness areas — diet, fitness, sleep, mental health, and habits.

Let’s not treat our users as objects but as co-designers. Take data usage: What does consent look like in a world where we’re co-creating? How might we represent everyone’s diverse needs and perspectives? Compliance and regulations are slow. The Golden Rule still applies: it’s up to us to do what’s right.

When the Metalab team works on AI products, we take a mindful approach to designing for permissions and privacy. We aim to be transparent with users about the data the product needs in order to reach its potential. After all, we can’t bring value unless we have enough data to draw patterns and be proactive, so it’s a constant line to walk in product design.

For example, Adobe, Getty and Shutterstock only use licensed and publicly available data to train its AI models. It also compensates creators whose work is used for the training process.

Another way to bring users into the process is by involving them directly in product development. For example, we worked with The Atlantic, a multi-platform publisher, to create The Atlantic Sandbox for AI testing and iteration — explore the 

This R&D site allows The Atlantic team to share experiments publicly without affecting their core product and environment. It showcases its rich catalog of 167 years of journalism, engages readers in feedback, and strengthens relationships with early adopters eager to trial new technologies.

Another example I’m watching is Jasper AI, a writing assistant that helps content creators. It works like a partner to generate ideas, tackle writer's block, and streamline the writing process. Jasper is designed to work alongside human writers rather than replace them. There’s a ton of interesting “assistant like” tools emerging, and those who act like a partner with great UX will be the next unicorn.

What kind of relationship do you want to be in?

AI excels at handling the tedious and the predictable, but it lacks novelty. It can't create new insights or styles beyond its training. It can’t feel. As AI becomes ubiquitous, our exclusively human qualities — our intuition, emotions, and imagination — will become even more valuable.

As makers, our job is to sense new possibilities and imagine things that are not yet conceived. If we want this motorcycle of the mind to head in the right direction, realize our actions have world-changing implications.

We can make things that are toxic to the relationship.

We can design and build with our intuition and humanity. We’re humans creating for other humans, and right now, the choice is ours. I think we should own it!

Designing a New Relationship with AI