Human-centered design was the through-line at Hawaiʻi Tech Week

Sep 15, 2026

Key Takeaways

  • Human-centered design has a second half. The first is building what people want. The second is refusing to cause harm.
  • AI is good at production and bad at knowing what is worth producing. Judgment is what marketers are paid for now.
  • Third places are back. In-person connection came up in nearly every HTW session I attended.

Ovetta Sampson spoke on a topic during Hawaii Tech Week that kicked off an underlying theme I sensed throughout the event. 

Human-centered AI design is Ovetta’s area of expertise, and it has two parts. The first half is building things people want. The second half, which many tend to skip, is declining to build things that harm people — cognitively, psychologically, or physiologically.

That second half is where AI gets interesting for marketers. AI is good at production and bad at knowing what is worth producing. 

I’m only referencing three hosts, three formats, one week, and a fair amount of overlap in who showed up, so I am not claiming a scientific sample. But it struck me that an AI executive, a creative leader, and a group of local marketers kept circling the same thing without prompting. People are what matters. And right now people want connection, cultural context, and some sign that a human made the thing.

 

Ovetta Sampson speaks into a microphone during a fireside chat with Gidens founder and CEO Vincent Brathwaite at the Inspiration Hawaiʻi Museum during Hawaiʻi Tech Week.

Ovetta Sampson in conversation with Gidens founder and CEO Vincent Brathwaite at the Inspiration Hawaiʻi Museum.

Ovetta Sampson on what design owes people

Gidens hosted Designing the Future with Ovetta Sampson at the Inspiration Hawaiʻi Museum, a fireside chat moderated by Gidens founder and CEO Vincent Brathwaite. Ovetta was candid and often blunt, and she clearly enjoys teaching what she’s learned through building.

She founded Right AI and previously served as director of machine learning UX at Google and VP of ML experience design and responsible AI at Capital One, with earlier roles at Microsoft and IDEO. She has spent more than 15 years exploring where design meets artificial intelligence.

Her line on adoption landed with the audience: AI is like ketchup. You can put it on anything, but it does not belong on everything. She was direct about what the work requires. “We do user research to be less wrong. Not to be right, but to be less wrong.” That is a humbler standard than most marketing teams hold themselves to, and a more useful one.

She was equally direct about how designers should show up. “We’re not order takers. We’re not waitresses. We’re collaborators.” Swap in marketers, and the sentence still works. The teams getting real value from new tools are the ones setting the guidelines.

In a recent column for ACM Interactions, Ovetta describes exerting control as making sure the outcomes machines produce are not harming people, and calls governance a nonnegotiable part of the instructions we give them. She frames the work as a loop rather than a handoff, with a person setting context and reviewing what comes back at every turn. The article is free to read and worth your time.

 

Dan Kim, a creative leader at IDEO, speaks during a Futuring Mixer at PEWA by Pono Potions, with a projection of illustrated Future Signals cards behind him.

Dan Kim of IDEO walks through Future Signals at the Futuring Mixer, hosted with UXHI at PEWA by Pono Potions.

Dan Kim on designing for connection

Dan Kim, a creative leader at IDEO, hosted a Futuring Mixer at PEWA by Pono Potions with UXHI. The event featured what he calls Future Signals, illustrated cards projected on the wall and laid out like a tarot spread, each one describing a shift already underway. We sorted ourselves as each card came up, stepping left if the change was already hitting our industry and right if it was not.

The cards that gave me hope were the ones about physical space. Third places are coming back, and some segments are choosing less screen time while the appetite for wearable body metrics also appears to be cooling. Retail is being redesigned around gathering rather than transacting, which raises a question I did not expect to be asking in 2026. Could malls make a comeback? My 13-year-old self certainly hopes so.

There is a reason people want somewhere to go. The U.S. Surgeon General issued an advisory in 2023 describing loneliness and isolation as a public health crisis affecting about half of American adults. If a nation is feeling short on social connection and tired of screens, then the live event, the store, and the room full of people become premium products.

 

For marketers,  judgment is the deliverable

The third session was ours, in the loosest sense of the word. Marketers in Hawaiʻi is an informal peer network my friend Jaclyn Loo co-founded. This year, we co-hosted Coffee, Chat & AI Strategy, bringing together roughly 35 marketers, and I moderated a panel with Avery Cavanah of ProService Hawaii, Erin Khan of Anthology FINN Partners, Brent Shiratori of Aidia Marketing, and Kristine Stebbins of Bank of Hawaii.

Brent had the sharpest version of why human judgment still matters. “AI has never lived a day in its life,” he said. “It doesn’t actually know what it’s like to actually taste a cup of coffee. It just doesn’t.” It can only describe it.

 

Three examples from the hour made that concrete:

  • Brent had a client from outside Hawaiʻi whose AI-generated copy came back stuffed with Hawaiian words that did not make sense and could read as offensive
  • Avery killed an AI-generated image of a local boardroom because the people in it were fake
  • Erin ran storm messaging for a tourism client through a framework to decide when to keep posting and when to pull back

Kristine markets a Hawaiʻi brand that is nearly 130 years old and said the bank is a long way from turning cultural context over to a machine. Her larger point stuck with the audience: “The paradox is now, as we get even more technology, the human and psychology aspect is going to become even more important.”

Five marketers pose together during a Hawaiʻi Tech Week event, wearing leis in front of screens displaying a “Marketers in Hawaiʻi: Coffee, Chat & AI Strategy” presentation.

The Marketers in Hawaiʻi panel at Entrepreneurs Sandbox, Sept. 3. From left: Tiffany Schaar, Avery Cavanah, Erin Khan, Brent Shiratori and Kristine Stebbins.

Brent also made a point about credibility that research supports from a different angle. Hand him an obviously AI-generated proposal, and he loses confidence in the sender, because the tells are easy to spot now. A study published last year in Organizational Behavior and Human Decision Processes ran 13 experiments and found that people who disclosed using AI were trusted less than people who said nothing. Trust fell furthest when someone else revealed the AI use for them. Audiences are already suspicious and disclosure can hurt, but getting caught costs the most.

The economics shifted underneath all of this. Brent used to get paid for making ads and buying media. Production costs very little now, so clients pay him for ideas. At his company, three people serve seven to eight clients, which he said would not have been workable a few years ago.

When we asked what delivered the best return this year, in-person events came up twice, alongside partnerships. People wrote answers on Post-its while standing together at an in-person event.

 

Both halves of a human-centric question

What I’m taking away from the week is not a workflow change. It is Ovetta’s two questions concerning human-centric design, which I have started running on my own work: Does this serve the person on the other end? And could it hurt them?

Marketing has gotten good at the first question. We research, we test, we optimize for what people say they want. The second question rarely makes it into a scope of work, and it determines whether what we ship was worth building.

I do not have this fully worked out. After experiencing a week of sessions showcasing unique perspectives, I know the answer will not come from an LLM. It will come from people with lived experience, who know what a word means to the community that owns it, and who can and will say “no” when a project is not the right thing to build. 

—Tiffany Schaar