Great AI will be defined by how it listens

AI is almost everywhere. It writes, draws, codes, plans, and executes. It provides insights, recommendations, and sometimes even life advice. While this is impressive, it prompts reflection on what kind of design AI truly needs and the nature of our relationship with it.

From a designer’s perspective, current AI systems are heavily optimized for speed. They're engineered to deliver quick answers, summaries, bullet points, and task completions. They function like supercharged search engines or ultra-efficient assistants. This efficiency is valuable, but it often narrows the space for exploration, reflection, or even disagreement.

Most of today's AI use cases focus on efficiency and time-saving. With the exception of medical applications, where the stakes and goals are quite different, the dominant narrative is about completing tasks faster and reducing friction. That’s understandable, but I hope this changes over time. Because what will truly take us to the next level is not just faster execution, it's new perspectives and new ideas. We should avoid fast-food version of productivity. Quick, convenient, efficient. No fries, no joy, just speed.

Creativity lives in friction, not in flow

The most enriching conversations I've experienced weren't efficient. They were messy, unfolding with pauses, contradictions, and ideas that didn't always lead somewhere but sometimes did. Today's AI isn't optimized for such interactions. Yet.

There’s a risk that by focusing solely on output and speed, we forget that creativity often emerges from friction, not flow. The pressure to reduce cognitive effort and eliminate ambiguity can be helpful, but it can also flatten everything that makes the creative process meaningful. We need to be careful not to design AI that smooths over the very bumps where insight lives.

Instead of thinking of AI as something that just responds, maybe we should be designing it to be very context-aware. To know when to give answers and when to ask better questions. When to be fast and when to slow us down. That balance is where collaboration actually happens.

Will AI always wait to be asked?

Right now, AI is something we prompt. We ask the question. We set the tone. We provide the context. And it responds, often instantly and with surprising coherence. But the quality of what we get still depends almost entirely on how well we ask.

This raises an important question: will AI always wait for us to speak first? Or will it learn to participate more like a human would, noticing gaps, asking questions, suggesting directions?

An AI that becomes proactive could shift the dynamic entirely. It could notice when we're defaulting to familiar choices or when a missing variable might matter more than the ones we're tracking. That sounds like a small thing. But it isn’t.

In real collaboration, we don’t just value answers. We value people who challenge us, reframe the question, or offer a new lens. If AI can step into that space thoughtfully, it stops being a passive tool and becomes part of the thinking process itself.

Real intelligence knows when to speak up. To do that, AI needs to act more like a teammate than a tool. It should be able to sense when we're stuck, when we're rushing, or when we're about to fall into the trap of obvious decisions. It should help us stretch our thinking, not just confirm it.

But with that comes responsibility. Proactive AI requires careful design. Not just in what it does, but in how it knows when to step in and when to stay silent. In some moments, we want flow. In others, we want friction. AI needs to understand the difference.

And maybe that’s where design plays its biggest role yet, not just shaping what AI looks like or how it talks, but how it listens. How it chooses when to act. And how it learns to be more than just efficient.