In April 2023, I was asked, “Grandpa, did you know that your teacher cannot now tell if you or an AI chatbot has written an essay?” I did not. I was flabbergasted to discover that, broadly, it was true. Moreover, improvements appeared to be accelerating.
I started to use the “free” Chat GPT-3.5; I have reviewed the answers to 234 of my questions to date. Within the answers, I have detected (and corrected as requested) just one “hallucination”. I realise that the sample size is tiny but it is merely a “real-world” illustration. I have been keeping an eye on the rapid “evolution” of AI, language chatbots, artists’ facilitators, deep fakes and so on for about eight years.
The PJ feature about pharmacy and AI by Christine Parry is timely and important. My take is that an AI may simulate or mimic awareness like a loquacious parrot but there will always need to be a human in the loop.
In pharmacy, the human, presumably, will be a registered pharmacist using a high-quality, curated database, such as the latest MedicinesComplete. That individual pharmacist will be privileged by experience of face-to-face personal contact with patients so, presumably, will be empathetically aware of their vulnerabilities. That pharmacist will take personal responsibility, guided by pharmaceutical morality and ethics. And, crucially, that human healthcare professional can also be held to personal account and could, ultimately, be struck off.
This chimes with the indeterminacy/technicality (I/T) ratio developed by sociologists Jamous and Pelliolle: technical knowledge is found in books; indeterminate knowledge is discovered during practice and is seldom or never in books. Higher-prestige professions, such as in the church, law and medicine, have higher I/T ratios.
The AI industry itself has asked for more regulation. Two of the three “godfathers” of AI have chosen to overtly warn that AI may have consequences on par with a pandemic or nuclear war. One of those “godfathers”, Geoffrey Hinton, warns that AI could very rapidly become more intelligent than humans. One reason is that communication of enhancements in AI is “instantaneous”, instead of the deep time demanded by biological evolution. Some humans have always attempted to control other humans; AI may help.
Finally, arguably, it is now pointless to learn anything, including about pharmacy, as so much human knowledge, certainly of type T and, arguably, at least a trace of type I, is now “instantly” available by swiping a computer screen.
Within swathes of scholarship, including all languages, literature and history, and occupations such as pharmacy and medicine, what is deemed “true” and “false” and much else may be about to experience unprecedented turbulence.
Of course, such extreme scenarios may be hype, conspiracy theory or just the latest moral panic, such as those surrounding drugs, such as LSD, marijuana and methadone, and novel technology, such as television, video games and smartphones. Within the latter, this is just the latest iteration.
But I suggest that it would be prudent for pharmacists and others to keep their eyes wide open.
Malcolm Brown, retired pharmacist, sociologist and author of Pharmacy Miscellany: Thoughts on regulation, education, industry and philosophy.
- 1Jamous H, Peloille B. Professions or self-perpetuating systems? Changes in the French University Hospital System. In: Jackson J, ed. Professions and Professionalization. Cambridge University Press 1970:110–152.