NHS doctor and digital health advisor, Dr. Annabelle Painter reveals her love of science and painting plus the challenges of preparing the clinical workforce for the introduction of AI technology



Dr. Annabelle Painter is a NHS doctor and a digital health advisor. She studied Medicine at Oxford medical school and has a degree in Natural Sciences from the University of Cambridge. She worked at Babylon Health as an AI Clinician where she deployed a triage chatbot to an A+E setting in several NHS trusts and led the development of a COVID-19 symptom checker.

Annabelle continues to practice as a GP registrar alongside working as an advisor to healthtech start-ups. In August, she will be starting an ‘AI & Workforce’ fellowship at HEE and NHS X which  focuses on preparing the healthcare workforce for the introduction of AI technology. She sits on the Royal Society of medicine Digital Health council and co-founded the Digital Clinicians Network – a community for healthcare professionals working within digital health.



What initially sparked your interest in medicine, and subsequently, AI in medicine?

I have always been fascinated by science and biology but I was unsure if I wanted to do medicine. That’s why my initial degree was in natural sciences at Cambridge University. It was during my undergraduate years that I began to realize I am into human biology and want to have a career where I can help people and apply that domain knowledge practically. That’s when I decided to pursue medicine.

When I was working as a junior doctor, I was frustrated at how technology limited me from delivering the best care. Technology within the NHS has been neglected for a long time and many of us are still working with paper notes, fax machines and pagers today. As someone who grew up with technology, it’s vexing that I cannot leverage the same amount at work. That drives me to make a difference, I want to change and bring in the kind of technology that we need in the clinical space.

What were your responsibilities as an AI Clinician at Babylon Health?

Before going back to clinical training, I was at Babylon Health for over a year working as an AI Fellow. My role was to provide clinical knowledge for the team during decision-making, product planning and revisions while products were being designed. I was also involved in the clinical validation and post-market safety reviews of products.

Specifically, I helped in the development of a chatbot triaging symptom checker to be deployed in the A&E setting. A huge percentage of people who came to the A&E department could have been attended by GPs or pharmacists so the aim was to advise people where best to get help for specific problems. I spent a lot of my time working on that until towards the end of my term when COVID-19 started.

We created a COVID-19 symptom checker which was a challenge since the pandemic was changing continuously and the government was constantly updating the guidelines at the beginning. Nonetheless, it was rewarding too as you feel like you are contributing something to help.

Do you think AI is prone to aggravating gender and health disparities?

I agree there is a risk. Like any tool, it is about how we use it. In terms of health inequalities, I think it is worth bearing in mind that the young, fit and healthy are most likely to use AI. Whereas those who may be older and less technologically literate tend to have higher healthcare needs. So, we need to be careful and prudent about how we introduce AI, so that we will not leave behind certain groups of people.

For gender inequality, I think it’s important to look at our datasets and think carefully about how they’d be used in healthcare and the possible inherent biases when it comes to diagnoses. It does not mean that a man showing symptoms associated with a condition that is more prevalent in women, we can overlook the possibility of him having that condition. We need to ensure AI does not accentuate existing biases.

It is also crucial to consider representation. If certain ethnic groups are not represented in the dataset, we may end up with an AI model that obsoletes them. Some diagnostic AI disregarded individuals with darker skin color and this resulted in inaccuracies. I think we need to address all these in the datasets to avoid people being disproportionately underserved by the technology.

Did you ever consider a different career?

Yes, I remember exploring lots of different career options when I was doing my natural sciences degree. I got a first-class after my first year in Cambridge and so applied to 20 consulting and pharmaceutical companies as a summer intern but got rejected by every single one of them without an interview. It was disheartening but it was also a good introduction to the experience of failure. I went traveling in the end and had a great summer. Looking back, if I were successful in my application and ended up working for one company that summer, I would probably not be doing medicine.

The other time I could have had a different career was making the difficult decision to go back into clinical training. The route of doctors wanting to work with AI and digital health had not yet been established. They were pretty much new areas without a standard career pathway. At one point, I found it hard to carry on because being a doctor and seeing patients is especially important to me. I don’t want to lose them in the process of pursuing AI and digital health. So I am glad that I am now working towards a career where I can do a bit of both.

You are about to start a fellowship focusing on AI and the workforce. What are some of the challenges of getting AI into the clinical workforce?

One of the main challenges is trust; how do we get the workforce to trust AI and use it in a way that they will feel confident about it and regard it as a means to improve efficiency and outcomes. We need two things to endorse trust. One is clinical validation. We need to be able to prove that AI works, like the way we prove a new surgical technique or drug works. Next is education. We need to evaluate the technology and access its effectiveness from time to time.

As such, AI becomes a whole new knowledge base and skillset that we need to disseminate across the healthcare workforce, in a way that’s understandable. It’s a huge challenge because there will always be resistance to change. We don’t want to tell the workforce what to do; we want to integrate AI like a collaborative exercise so we can make the best use of the technology. I believe my fellowship is going to be interesting; exploring all the challenges and navigating ways to overcome them.

What advice would you give someone starting their career in medicine or medical AI?

If you’re a medical student or junior doctor and would like to work in digital health or AI in the future, the best thing to do is to get involved in related projects. Regardless of how small they are, just get your hands dirty and build up your experience in the areas. Having certain skillsets, like data science, statistics and some basic knowledge of coding will be helpful. It’s important to build these skillsets alongside your clinical training.

There are many things you can do to show your interest in AI. The easiest way is to attend conferences and have a sense of what is going on in the field. Don’t be afraid to be unconventional because everyone’s experiences and career goals are different. Medicine has always had this very strict route and stepping away from it can be extremely uncomfortable. Some people may even question why you are doing that. However, sidetracking from your usual clinical training is going to make you stand out from the crowd.

It’s important to encourage medical students and junior doctors to try out different things so that our medical workforce will have a much more diverse skillset and have more to offer.

What’s the best piece of advice you ever received?

Don’t rush to complete your clinical training. I heard it from many different people, and it’s such important advice. I realized recently that you can spend a lifetime climbing a ladder, only to realize that it was up against the wrong wall. I think a lot of people rush to the top or an end goal and then, they look back and think, ‘Why didn’t I do all those other things that were there?’ I believe it’s more important to diversify your skillset and experiences rather than just rush to the top.

How do you relax?

I’ve painted since I was a child and it’s the only time I switch off. It’s real escapism. I think that kind of creativity helped get me interested in AI and digital health because I like to come up with innovative solutions to overcome challenges. I’ve combined my love for science, problem-solving, arts and creativity to get me to where I am today.

Given your surname, were you born to paint?

It’s funny because no one else in my family paints and no one else does anything artistic. Maybe it’s coincidental but when I was young I remember thinking being called ‘A Painter’ would be perfect when it comes to signing my paintings!

How do you balance your everyday work and your love for the arts?

It’s difficult. I have clinical training, I also advise several startups and manage digital clinicians that work with co-founders. But like most people, they try to find time to do things that they love, be it cooking, watching Netflix, or meeting up with friends. I’m the same – I try to fit in painting whenever I can. I also try to incorporate AI and digital health into my practice whenever I can. The pandemic lockdown has given me more free time to do the things I love, whereas normally, I wouldn’t have had that luxury. As the saying goes, you can’t do everything, you can only prioritize.


Dr Painter is a speaker for our Clinician Series focusing on primary care and population health. See full details and register for free here.