Creative Comment | We have been here before

The greatest opportunity AI presents may be to remind us what only humans can do.

Earlier this week, I visited a dear family friend living with late-stage Parkinson’s disease and dementia. Conversation has become increasingly difficult. Memories appear and disappear unpredictably. Much of the woman I’ve known for decades has gradually slipped beyond reach. We sat quietly together for a while, holding hands, until she looked directly at me with remarkable clarity and said something I haven’t been able to stop thinking about:

“You are a human being.”

The sentence was almost absurdly obvious. Of course I’m a human being. Yet in that moment, it felt less like an observation than a reminder. I’ve carried those five words with me ever since, not because they revealed something I didn’t know, but because they reminded me of something our industry seems increasingly willing to forget.

That realization stayed with me throughout a week dominated by conversations about artificial intelligence: synthetic users, digital personas, virtual ethnography, machine-generated empathy, and AI systems that promise to understand people faster, cheaper, and at unprecedented scale. Somewhere between those conversations and my visit with my friend, I realized I was witnessing the second great technological revolution of my career. More importantly, I realized it was asking us the same question as the first.

More than 20 years ago, I began my career at one of the most reputable agencies that helped establish user experience as a business discipline. At the time, the internet was transforming business, and organizations were captivated by technological possibility. Every conversation revolved around what digital products could do. New capabilities dictated product roadmaps. Engineering drove experience. If the technology existed, someone was already trying to commercialize it.

The agency I worked at represented a fundamentally different philosophy. Instead of beginning with technology, we began with people. We spent time observing how customers actually lived, how they navigated websites, adapted products to fit their routines, hesitated before making decisions, and quietly invented workarounds that no one had anticipated. We weren’t simply validating requirements or confirming assumptions. We were looking for the moments that challenged them.

That shift changed the trajectory of digital. User experience wasn’t simply another design discipline; it was a correction. It reminded an industry captivated by technology that innovation succeeds only when it begins with the lived experience of another human being. We stopped asking, What can technology do? and started asking, What does another human being need?

Looking back, that lesson feels self-evident. At the time, it was revolutionary.

Today, as artificial intelligence reshapes nearly every aspect of business, I can’t help wondering whether we’re quietly drifting back toward the same technology-first mindset that UX emerged to correct.

To be clear, I am not skeptical of AI. Quite the opposite. I started an AI consultancy three years ago. I fundamentally believe that AI has transformed how I think, write, synthesize information, and solve problems. Like many people, I believe it will become one of the most important productivity tools ever created. The question isn’t whether AI will change research. It already has. The question is whether our excitement about its capabilities causes us to misunderstand the purpose of research itself.

Research has never existed simply to collect answers. It exists to discover questions we didn’t know enough to ask.

One of the first lessons every researcher learns is that people are remarkably unreliable narrators of their own behavior. Ask someone how they grocery shop, then accompany them to the supermarket. Ask an executive how strategic decisions are made, then quietly observe the executive meeting. Ask a patient how they manage a chronic illness, then spend an hour in their kitchen.

The distance between what people say and what they actually do is not evidence of dishonesty. It is evidence of being human. Much of our behavior operates below conscious awareness. We rationalize decisions after we’ve made them. We adapt to friction until we stop noticing it. We invent rituals, shortcuts, and workarounds without recognizing that we’ve done so. The work of a researcher has never been simply to gather information. It has been to notice what people themselves cannot yet articulate.

This is where I believe much of today’s conversation around synthetic research becomes confused. Large language models are extraordinary pattern-recognition systems. They synthesize information at a scale no individual researcher ever could. They identify relationships hidden within enormous datasets, generate hypotheses, summarize interviews, and surface inconsistencies that would otherwise remain invisible. Those capabilities are genuinely transformative, and every serious research practice should embrace them.

But synthesis is not discovery, and prediction is not understanding. A model predicts what is most likely based on everything it has already seen. Discovery often begins with the thing no one has seen before or thought to ask.

Nearly every meaningful breakthrough I’ve witnessed over the course of my career began with an observation that surprised everyone involved. Sometimes those moments were absurdly obvious. “How did we not realize this?” Those moments don’t emerge because someone asked a better question. They emerge because someone paid closer attention and stopped over-engineering.

Technology has always tempted us toward abstraction. We reduce people to users, customers, audiences, market segments, personas, and now synthetic humans. Each abstraction makes decision-making more efficient. Each one also risks placing another layer between us and the people we claim to serve. Humanism pushes in the opposite direction. It insists that before we understand a market, we understand a person. Before we trust a model, we spend time in the real world. Before we assume we know the answer, we remain curious enough to be surprised.

I’ve returned often this week to my friend’s words: “You are a human being.” In a week filled with conversations about synthetic humans, artificial empathy, and machine-generated insight, someone whose own memories are slowly disappearing reminded me of the one thing all of this technology is ultimately meant to serve: another human being.

I’ve found myself repeating one sentence ever since:

Empathy is not pattern recognition. Empathy is the willingness to be surprised by another human being.

Too many organizations are approaching AI as a substitute for human insight rather than an amplifier of it. They begin with the model and ask whether people still need to be involved.

I would argue for the opposite sequence.

• Begin with people.

• Use AI to make sense of what you learn.

• Return to people to test, refine, and challenge your conclusions.

In other words, let AI accelerate the science of research while protecting the humanity of it.

The organizations that win in the next decade won’t necessarily be those with the most sophisticated models. Those capabilities will become increasingly accessible to everyone. The advantage will belong to companies that build cultures of observation — organizations that remain relentlessly curious about customers, employees, and communities.

The future isn’t AI-first. It’s AI-first, human-led.

We are, after all, human beings.

Featured in GDUSA.