In my colleague Ann Somers Hogg’s recent blog, she argues that Jobs to Be Done Theory is a good starting point to evaluate gen-AI’s use in health care. I had a similar thought, and remembered a tweet I had seen online by Joanna Maciejewska, a fantasy and sci-fi author. It said: “You know what the biggest problem with pushing all-things-AI is? Wrong direction. I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”
I understand the excitement around AI: ChatGPT is intriguing, and so are AI art generators. But are they helping people accomplish their “Jobs to Be Done,” a.k.a. their desired progress in specific situations?
Each day “jobs” arise in our lives, and we “hire” products or services to accomplish these jobs. Understanding an individual’s job is powerful because it enables innovators to develop a product or service that aligns with what someone is already trying to accomplish.
This is part of Jobs Theory or Jobs To Be Done Theory, a framework that helps explain the causal drivers of behavior, developed by Clayton Christensen and Bob Moesta.
Organizations should be designing AI to help people accomplish their jobs. This way, a service or product helps people while gaining consumers and growing its business. Otherwise, you risk developing a product or service that attracts attention for a moment because it’s new and different…but not sustainable.
Putting AI to good use: supporting autism diagnoses
Applying this logic, one area in which AI could help serve unmet Jobs to Be Done is within behavioral health. Specifically, AI could support autism diagnoses.
In my prior role as an attorney, I used to represent parents on behalf of their children with unmet mental health needs. Several of my clients thought their children had Autism Spectrum Disorder (ASD), but they didn’t have a diagnosis and couldn’t get necessary supports.
Getting an ASD diagnosis in today’s health care environment is challenging. There’s no simple medical test like a blood test to diagnose ASD. Therefore, to diagnose a child, a doctor will look at the child’s developmental history and behavior.
Having an ASD diagnosis as early as possible is key to getting a child the services and support they need to thrive, and yet it requires an assessment by a professional trained and experienced in diagnosing the disorder, such as a pediatric psychiatrist, neurologist, or developmental pediatrician. And these days, those are in short supply.
Another challenge is that these professional assessments aren’t perfect. Diagnoses are based on the child’s behavior during a finite period of time plus their parents’ observations of their child’s behavior. As a result, these circumstances may not be the most reliable sources of information for accurate diagnoses.
And then there’s the fact that these assessments are costly, time-consuming, and providers vary on the insurance they will accept, if they accept any at all.
Helping solve some of these challenges is where AI could help.
Children could more easily receive ASD diagnoses and AI could make diagnosing more accessible. It could provide a solution to parents’ potential job of “when I am concerned about my child and getting a diagnosis is too difficult because of the time, expertise, and/or money needed, help me get my child assessed, so that I can get her the support she needs as soon as possible.”
But is this scenario merely a fantasy? It may not be so far-fetched.
Already, in using an electroretinogram (ERG), researchers have demonstrated the feasibility of using AI to identify specific features to detect ASD. An ERG is a “diagnostic test that measures the electrical activity of the retina in response to a light stimulus.”
The researchers used an ERG to measure retinal responses of 217 children 5-16-years-old. Some had been diagnosed with ASD, others did not. The researchers found that the ERG generated a different retinal response for the children with ASD as compared to the children who didn’t have ASD. The test takes as little as 10 minutes.
They believe that further research could provide clinicians with an improved, non-invasive method for diagnosing ASD, fast-tracking support for children with ASD, and alleviating time, stress, and costs for their families.
There is a lot of hype around AI right now, but if companies aren’t seeking to develop products or services based on people’s jobs, they risk a quick rise and fall.
Instead, when developing AI products, companies should be focusing on people’s desires for progress in their current circumstances. To do so, they should conduct a Jobs analysis. If you’re interested in what that would look like, check out this blog by Ann Somers Hogg.
