OpenAI dropped ChatGPT Health earlier this year. It promised to gather your scattered medical data. To put it all in one place. It sounds convenient for people who already live on the internet.
The US burden of chronic disease hits underserved communities the hardest. These are the exact groups least likely to use new digital tools. So, who is this innovation really for? It isn’t built with them in mind. It is marketed elsewhere.
When the Assistant Hallucinates
Sergei Polevikov is a tech expert. Founder of AI Health Uncut. He uploaded his records to see how the platform worked.
It wasn’t smooth.
The system hallucinated parts of his medical history. He hit administrative walls just trying to log in. Polevikov summed it up bluntly. He gets the efficiency gain. He saves time. But he doesn’t need an AI to tell him what his body is doing.
“The distribution of these tools is one-sided and people like me, who are experts, need them the least.”
If an expert struggles, think about the rest.
Polevikov navigates roadblocks easily enough. But what happens to those with low digital literacy? Those whose voices were absent from the design meetings?
Tech-Ignored, Not Averse
I went to Mobile, Alabama, recently. Talked to people on the street.
The gap is stark. Leevonis Fisher leads the Bay Area Women Coalition. When asked about AI, her response was instant.
“I think it’s all fake.”
She owns an iPhone. She uses Siri. Alexa. Voice-to-text. Almost every single day. Yet, she has no idea those features rely on the technology she distrusts. To Fisher, AI isn’t a helper. It is a buzzword for deceptive computer robots generating fake videos.
This isn’t just a misunderstanding. It’s a design failure. Engineers build for efficiency. They build for users who don’t really need saving. Meanwhile, those who could benefit are left out. Fisher doesn’t know she’s using AI because the marketing didn’t speak her language. She isn’t part of the design conversation.
She won’t buy the wearables. She won’t log into the portals.
We tell ourselves these folks are tech-averse. They aren’t. They are tech-ignored. No one invited them in. No one showed them the utility. Until we fix that, the gap widens.
And let’s be real about the economics. Missing underserved markets is bad for business too. The full potential of health tech dies if it only serves the healthiest people with the highest literacy levels.
Are we actually okay with leaving millions behind because the interface isn’t intuitive enough?
Who Steers the Ship
It’s not a lost cause. Money is moving.
Last fall, the MacArthur Foundation joined Omidyar Network. Along with eight other groups. They launched Humanity AI.
A five-year initiative. Five hundred million dollars. The goal is clear: AI must be shaped by people, for people. Michele Jawando of Omidyar Network puts it simply.
“The future will not be written by algorithms. It will be written by people.”
That’s a strong start.
But consensus is slow. We need more than pledges. We need context. We need to listen to the people we keep pushing to the margins of the conversation. Until then, the tool works for the few who already knew how to use it. The rest of us wait.
