When Expertise Can Travel
The Wrong Kind of Luck
I recently heard a story about my sister-in-law, who used to teach special education, helping a mother and her daughter get what they needed from a school.
She knew the system. She knew the language. She knew how to ask the right questions without immediately putting everyone on the defensive. She understood not just what the child needed, but how institutions often work: what gets said plainly, what gets softened, what gets delayed, what gets offered as kindness instead of recognized as a right.
That kind of knowledge matters. In the world as it actually exists, it can change the shape of a child’s education.
What stayed with me was not only that she was good at it. It was how rare that kind of help really is. Most people do not have a special education expert in the family. Most people do not have someone who can translate a school meeting into plain English, or tell them when a vague answer is merely vague and when it is quietly evading an obligation. A lot of people walk into rooms like that tired, worried, and already half-convinced they are not qualified to push back.
I do not hear a story like that and think, good, let’s replace the expert.
I hear it and think: this is one of the reasons I feel real optimism about AI.
Not because I think a machine can become better than people like my sister-in-law. Not because judgment, care, patience, or advocacy have suddenly become optional. People like her are indispensable. The problem is that indispensable people are not evenly distributed. Their knowledge is often trapped inside geography, income, timing, networks, and luck.
That is the wrong kind of luck to build a life around.
The Maze Learns Your Name
Part of why this lands so hard for me is that I have lived a version of it myself.
I stopped working about a year ago. Anyone who has to step out of work because of disability learns very quickly that the difficulty is not only the body, or the illness, or the grief. It is also the maze. The forms. The timelines. The clauses. The phone calls that answer one question by opening three more. The dense language that seems designed to keep ordinary people slightly off balance.
Suddenly I had to understand systems I had never wanted to become fluent in. SSDI. LTD. Medical records. Work history. Eligibility. Rights. Deadlines. What I was entitled to. What needed to be documented. What one decision might do to another.
AI was genuinely useful to me there.
Before I left work, it helped me weigh my options and think through the consequences. After I left, it was instrumental in helping me secure SSDI and LTD. Not by magically solving the process. Not by replacing doctors, records, official decisions, or actual human expertise. It helped by giving me a place to bring my confusion. It helped me ask better questions, interpret what I was reading, organize what mattered, and keep the bureaucracy from becoming one large, shapeless threat.
That may sound modest, but anyone who has dealt with disability systems, insurance systems, school systems, or legal systems knows that orientation is not a small thing. Sometimes a person does not need omniscience. They need a clearer next step. They need the sentence in front of them translated into something they can act on. They need to move from panic into motion.
That is where much of my optimism begins. Not in the fantasy that AI makes experts unnecessary, but in the much more human possibility that it lets more people borrow a little of the clarity experts carry around.
Where Help Fails to Arrive
This is bigger than my own life, of course. My story is one example of a larger pattern: expertise exists, but access to it is uneven.
You can see that in healthcare. The World Health Organization estimates a projected shortfall of 11 million health workers by 2030, mostly in low- and lower-middle-income countries, and notes that shortages are worsened by difficulty deploying workers to rural, remote, and underserved areas. In practice, that means knowledge may exist somewhere, but still fail to arrive where and when people need it.
That is part of why certain uses of AI in medicine strike me as meaningful. A 2022 Communications Medicine study looked at a mobile AI system for fetal ultrasound in under-resourced settings, using blind-sweep ultrasounds collected by sonographers in the United States and Zambia and by novice operators in Zambia. The system estimated gestational age at a level non-inferior to standard fetal biometry, ran on Android phones in under three seconds after a sweep, and could work without internet connectivity. That is not a machine replacing an obstetric specialist. It is a fragment of specialist capability traveling farther than it otherwise could.
That line between research and real use is already moving. In March 2026, Reuters reported that Butterfly Network received FDA clearance for an AI-powered ultrasound tool that estimates gestational age in under two minutes and is intended for settings such as emergency departments, rural clinics, and other low-resource environments. The company also said the tool had already been deployed in Malawi and Uganda.
A similar pattern appears in eye care. In Rwanda, the RAIDERS randomized controlled trial tested whether AI-supported diabetic retinopathy screening improved referral uptake. The important part there is not only that a system can screen. It is whether screening helps people reach the next step before preventable vision loss becomes irreversible.
The same problem appears in law. The Legal Services Corporation reports that low-income Americans in rural households did not receive any or enough legal help for 94 percent of their substantial civil legal problems, and its rural justice work names barriers like attorney shortages, limited internet access, distance, transportation, and the need for community-specific trust. The World Justice Project estimates that about 5 billion people globally have unmet justice needs, including people who cannot obtain justice for everyday problems or are excluded from the opportunities the law is supposed to provide.
Those numbers are enormous, but the lived reality inside them is familiar. People are trying to understand high-stakes systems while tired, underinformed, and mostly alone.
Optimism, Properly Sized
This is the version of AI that matters most to me.
Not the loud version. Not the version where every profession is supposedly five minutes from extinction. Not the version where every human bottleneck is treated as waste. I mean something smaller, steadier, and more humane.
I mean a parent walking into a school meeting with better questions and less fear.
I mean a disabled person staring at a wall of paperwork and not having to decode every sentence alone.
I mean someone in a rural town, a poorer country, or a place with too few specialists and too much distance getting a useful first layer of guidance instead of nothing.
None of this removes the need for experts. If anything, it reveals how badly we still need them. My sister-in-law still matters. Doctors still matter. Lawyers still matter. Advocates still matter. Teachers still matter. Human beings with experience and judgment still matter.
Still, scarcity is not the same thing as wisdom. Sometimes scarcity only means distribution is broken.
That may be the simplest way to say why I care about AI as much as I do. Not because I think expertise should disappear. Not because I think the best human capacities can be automated away. I care because so much of life is still shaped by whether the right knowledge happens to reach someone in time.
AI, at its best, can help some of that knowledge travel.
Less gated. Less local. Less dependent on whether you know the right person, have the right money, speak the right language, or happen to live in the right place.
That does not fix everything. It does not absolve institutions of responsibility. It does not turn injustice into justice or bureaucracy into mercy.
Still, it is not nothing.
Sometimes it is the difference between paralysis and motion. Sometimes it is the difference between hearing that help exists in theory and finding the first usable path toward it in practice.
That, to me, is reason for optimism.
Not because expertise should vanish.
Because it should travel.
People need more help
Expertise is essential
Now more can get it
