With the launch of ChatGPT, it now feels like we are truly in the age of AI. But we’ve been using it in many ways for some time, says Peter Kelly…

The list is long ... Spotify playlists, Amazon recommendation lists and so on. But now we all have access to a broad artificial intelligence (AI) tool as opposed to narrow AI. Broad AI is a term used to describe AI that is able to handle a range of tasks, while narrow AI is an application that is designed to handle a specific task.

The difference between the two is that broad AI is more versatile and can handle a wider range of tasks, while narrow AI is more focused and can only handle a specific task. ChatGPT is broad AI; the AI that makes Spotify playlists for you is narrow AI.

The first question everyone wants to know about AI is: ‘Will it take my job away?’ So I asked ChatGPT whether AI will replace pharmacists, and the good news is that it replied: ‘Not in the foreseeable future.’ ChatGPT does believe, however, that we will work together in unison very soon.

I then asked ChatGPT to write me an article for a pharmacy magazine. Is one job, the side hustle, safe for now? The article the software wrote was pretty good. Readable, interesting, if a little bland.

It wrote an article about the importance of medication compliance. It took about two seconds. It was plausible that a human could have written the article - and that made me realise the first wrecking ball of ChatGPT may well be school homework.

I have no doubt teenagers across the global are already signed up and chancing their arm by getting ChatGPT to do their homework. I would highly recommend playing around with it – the speed with which it dishes out a reply is quite extraordinary.

Broad AI today is where the sat-nav and GPS was 15 years ago. I remember when sat-navs first started to become widely used, every day on morning television there was another amusing anecdote about how a sat-nav said something non-sensical or led someone to drive into a field or river.

Similar stories about AI will come thick and fast over the next year or so. There is something comforting for us in the fallibility of new technologies but, like sat-navs, the errors will get ironed out in time and it will become much sharper.

This summer I am off to Edinburgh to perform comedy at the Edinburgh Fringe, so I asked ChatGPT to write me a summary of my show. It wrote me a lovely synopsis. It said the show starts with me attempting to get in shape in the gym and then I joke about how people use filters to look like animals on dating sites.

It all sounded great, the only issue being I have never joked or told stories about any of the topics the AI described me as doing, yet it described ‘my show’ with the confidence of a narcissist compulsive liar.

This could be the first challenge pharmacists will face with AI. Patients  will ask AI questions and get answers they take to be gospel and use those answers to question our knowledge or competence.

Stephen Colbert recently did a monologue on the Daily Show in America on AI where he mentioned a project where AI was fed loads of images of moles. Some were cancerous and some were not. AI’s job was to use the data to identify whether moles were malignant or not.

Many of the stock images of cancerous moles had rulers in the photos. AI came to the conclusion that rulers were malignant, so every time it was shown a photo of a mole with a ruler in the shot it assumed it was malignant as, statistically speaking, the data it was fed showed this to be true.

Personally, I’m excited by AI. I am slightly lazy by nature, and I love shortcuts. AI could be the ultimate tool for finding shortcuts to reduce workload. I think it will free us from many of the more mundane tasks and enable us to concentrate on the more creative tasks – and that applies not just to the pharmacy profession.

Society could benefit greatly if healthcare professionals like pharmacists and doctors had more time to consult with patients, particularly patients with mental health concerns.

Having said that, we are a long way off being able to trust the integrity of AI. AI applications will have to be observed closely for years to come and we should place trust in it only bit by bit. And we will have to be vigilant about bias within AI.

The racial and gender bias of technology is very real and will have to be tightly observed and regulated.

Peter Kelly is a community pharmacist based in London and a stand-up comedian.