One of the most salient discussions around AI is if it will lead to mass job losses and aggregate even more wealth in the hands of the few. I used the last 3 posts discussing AI agents and how the capabilities of these agents to process information, reason, and complete tasks is advancing. This will mean that many jobs that have historically been too complex for automation are now ripe for it. This could also include many healthcare jobs. For example, companies like Hinge Health and Sword Health have developed AI solutions that allow for AI-guided physical therapy. They use computer vision and agentic capabilities to monitor patient’s movements in performing exercises using a smart phone camera, analyze them, and provide feedback to the patient. This allows patients to do their physical therapy at home and receive guidance without the presence of a physical therapist. Does that, then, mean that physical therapists are now an endangered specie? The answer to that is rather complex and history can be a guide here.
As you can see in the graph, in a recent report, McKinsey projects that a high percentage of many common jobs will be automated in the near future, thanks to generative AI (gen AI.) Some of these roles, such as office support, can see as much as 87% of the activities automated. In healthcare, they estimate some 43% of the activities can be automated. This is a rather high percentage and probably spreads across a variety of jobs, including nurses, doctors, therapists, office staff, and more. While that sounds like a lot of impending layoffs and redundant staff, the realities of the state of the healthcare workforce tell a completely different story. As you can see in the graph below by Accenture, we’re facing a serious and worsening crisis in healthcare. There is a serious mismatch between the supply and demand of clinicians. This is projected to reach epic proportions by 2030. As illustrated in the graph, AI can address a big part of this issue as AI-enabled solutions are launched for different use cases. As you can see, even with AI, we face a shortage of resources. As such, it is highly unlikely that AI will lead to job losses. Rather, either the staff will do their jobs differently with the help of AI or they will be moved into other roles that are needed and can’t be automated with AI.
Given the fact that AI is a foundational technology and will touch every aspect of healthcare, it can assist this shortage across different roles. The graph below shows some of the projected use cases and their dollar value. It is not surprising that virtual nursing assistants and administrative workflows represent two of the biggest areas that AI will have an impact. The shortage of nurses and home aides is among the most pressing issues facing the healthcare industry. Also, the administrative burden of the US healthcare system is far higher than any other advanced economy. As such, being able to automate as much of it as possible will mean that those resources can re-directed to other short-staffed parts of healthcare that desperately need that help. There will need to be massive efforts undertaken to re-skill and up-skill the healthcare workforce. Those who continue to do the same jobs will need to be trained in how they will do those jobs with the assistance of AI. Those who will move into new roles will need to be trained for those. To minimize disruptions, there needs to be active planning for all of this at the national level.
We have been through all of this before in our history. In an upcoming article in Fast Company, I write about how very common jobs disappeared as new technologies became available. The examples I give are the farming jobs of the 1800’s, when more than 90% of the US population was in agriculture. Today, only about 1% of our population is in farming but we don’t have a high unemployment rate. Most of farming has been automated and mechanized and farmers are far more productive using these tools. Similarly, phone operator was one of the most common jobs in US in the 1950s but as telecom technologies advanced, there was no need for those roles. However, the telecom industry has millions more jobs than it did in the 1950s and they’re higher paying jobs. The nature of these jobs is more high-tech and the expanding capabilities of telecom means a variety of new roles.
We will need AI to come to the rescue as soon as possible as it is the only technology on the horizon that can provide some relief to our major shortage of human resources in healthcare at scale. My guess is that with the aging population, our needs for staff in a variety of roles will only get worse and we will need technology to alleviate these shortages. As such, the adoption of these technologies, including AI, will need to be accelerated to stay in front of this.
