One of my sons is a radiographer, working in a large public hospital, carrying some management responsibility while still being ‘on the tools’ in an under-resourced, bureaucratic and highly structured environment.
On one hand there is the health system, hobbled by rules, work practises inherited from another century, wrapped up in extreme risk aversion. On the other you have the doctors, ranging from the juniors who are hospital employees, to the specialists who after years of study and work have the opportunity to ‘cream’ the system.
Of interest here when considering the role of AI, is the relationship between the radiographers, who construct the images, and the specialist radiologists. The radiologists carry complete responsibility for the interpretation of those images, along with the directions for treatment to other medical branches that carry out the hands-on care of patients, from nursing to surgery. The radiographer just takes the ‘pictures’, and is prohibited from diagnosis, no matter how experienced they may be.
Being a commercial bloke, for years I have been asking my son, where to from here?
Being a public servant for life is not all that attractive to him, overworked, frustrated and grossly underpaid. On the other hand, to go into business for himself, the combination of the capital required for the imaging gear, and the simple fact that the regulations require that only a specialist radiologist interpret his ‘pictures’, means they have the private radiography game completely sewn up. No private radiography studio can set up without a Radiologist locked in to sign off every image.
However, AI is happening.
One of the earliest uses of AI has been to read medical images. Their ability to ‘learn’, and consistently improve means that the room left for interpretation by a human is being squeezed into an increasingly narrow field loosely described as, ‘So what now”. As this continues to evolve, the need for the specialist radiologist in diagnosis will disappear. With this increasing irrelevance, in a free market, my son could start his own radiography business. This should be free of the regulatory constraints that dictate diagnosis is only to be done by a Radiologist, whose role will be little more than to ‘sign off’ an AI generated diagnosis. Radiology is a medical speciality whose only role within a very short time will be answering the ‘so what now’ question, and that will be increasingly answered by AI, informed by the outcomes of previous cases.
I am sure the ‘Radiologist union’ will fight tooth and nail by lobbying, to prevent that from happening. They are a part of a very smart and very highly educated cohort who have made a huge investment of time and energy into their future, and are unlikely to easily let the rewards from that investment trickle away.
We have only just begun to think about the impact of AI in the wider strategic context, but it seems evident to me, just based on this small example, that huge changes are afoot, many of which will be hobbled by the past, making the changes necessary to leverage the capabilities of AI extraordinarily challenging.
