Once a year the Greeks celebrate the life and work of Hippocrates, 2500 years ago, who in many ways set the standards for medicine in the centuries that followed, and is memorialised in the Hippocratic Oath that all medical students affirm.
As part of the celebration, many students travel from around the world to Hippocrates’ island of Kos and take their oaths in an elaborate ceremony in his presence – or rather the presence of an actor making a passable impression of the good doctor. And senior Greek clinicians debate his legacy and influence.
This year I made my own contribution. I created what Claude calls a ‘tool’ to embody Hippocrates’ thinking, by training it on writings by and about the great man, and modern-day attempts to bring his principles up to date. The result is a program into which documents can be fed, to evaluate their alignment (or otherwise) with Hippocratic principles.
My first test was of the UK National Health Service’s 10-year plan. I was reassured to find good alignment on eight of the ten principles the tool had developed. However, in a couple of cases – assertions about genomics guiding care in a widespread fashion and (ironically) the NHS becoming the most AI-driven of systems – the tool picked up that the plan made these assertions without pointing out the challenges they would face and the limited evidence they would have the expected universally positive impact.
Given my interest in health longevity, and clinics offering these services, I could not resist testing a prospectus from that direction. Less impressive. Many promises with little evidence, as perhaps we might have expected.
As you will immediately realise, this approach could be adapted to any field in which we want to set standards, apply principles and assess alignment. Corporate HR policies and their application in recruitment, development and disciplinary practices. Local authority plans and statements and their enactment. Even political speeches, assessed against what might be regarded as values shared by most people in a country, or a set of religious principles, such as the Sermon on the Mount, the Torah, or the Koran.
So AI can assist leaders to set principles for documents and communications and test alignment of what flows in and from their organisations. As I understand it, lawyers are already doing this with the documents they review and the arguments they make. It doesn’t remove human accountability for truthful and helpful content but can help reinforce it.
It takes only a few minutes to develop tools like this, but of course if they are to be used in serious practice, they would need to be tested and the output reviewed by wise observers. But all that can be arranged.
I’d be interested to know where my readers would choose to take this approach, and what they learn when they do.
