Is your GP using ChatGPT?

How would you like to be seen by an AI doctor? You might already be seeing one. Indeed, it turns out that already one in five British GPs is using artificial-intelligence tools like ChatGPT in clinical practice. In the largest survey of its kind, I asked 1,000 GPs across the UK about their experiences with generative AI. Among the 20 per cent who have embraced these tools, nearly 30 per cent use them for filling in documentation after patient visits. Perhaps more alarmingly, for some at least, 28 per cent use AI to help generate a list of possible diagnoses, and 25 per cent use it to explore treatment options.

Bots such as ChatGPT, Google Gemini and Bing AI are the latest AI sensations. These generative-AI tools derive their responses from vast troves of data. Users can ask chatbots questions and, unlike search engines, they will deliver conversational responses. You can feed it a set of symptoms and it will spit out a list of potential ailments. It’s like talking to the internet, on steroids.

In public discussions and lectures I have given, doctors often express disbelief at my survey findings – that so many healthcare professionals are using AI to help them do their jobs. Yet privately, many others tell me that they have dabbled with these tools themselves. A study I conducted in the US in October 2023 – nearly a year after ChatGPT’s public release – found that four in 10 psychiatrists have used AI to assist with clinical tasks.

It’s not only doctors, either. AI is also encroaching on other fields. Last November, a Salesforce survey across 14 countries found that 64 per cent of workers had passed off generative AI’s work as their own. In February this year, a Pew Research study reported that a quarter of Americans use AI ‘almost constantly’ or several times a day. Younger people are the fastest adopters, with more than four in 10 people under the age of 30 having dabbled in ChatGPT.

The tech takeover of the workplace has been forecast for some time. In fact, AI, automation and robots have already come for many blue-collar industries, like automotive manufacturing and agriculture. Economists at MIT and Boston University studied the impact of robots on the US labour market from 1990 to 2007 and found that, on average, one robot replaced almost six human workers. This has had a ripple effect across the entire US economy. For every additional robot per thousand workers, the employment rate drops by 0.2 per cent and wages decrease by 0.42 per cent.

Economists and researchers working on AI and the future of work have warned for some time that we ignore these developments at our peril. But we must also recognise that change won’t happen overnight. As Oxford University economist Daniel Susskind argues in his book, A World Without Work, those waiting for the ‘dramatic technological big bang’ will be disappointed. Instead, the nature of our jobs will gradually shift and change.

In healthcare, education and other sectors, there is growing recognition that AI can disrupt how we work. The use of such tools certainly raises uncomfortable questions about plagiarism, responsibility and accountability. But at the same time, it would be foolish to panic about AI just yet. For starters, there is plenty that this tech cannot do. Consider the latest chatbots, which still struggle to understand basic sentences, follow simple instructions or grasp any kind of subtext. When asked for factual information, they can spit out erroneous but coherent answers – what experts call ‘hallucinations’. These bots are also not neutral when it comes to dispensing information. Rather, they reflect the biases of their human creators – though whether these are generally better or worse than those held by the humans themselves is still unclear.

It’s fair to say that nobody – not even the AI experts – really knows how or why the latest wave of AI tools can do what they do. Yet, as anyone who has used them can attest, they offer numerous benefits. And despite their current limitations, AI is constantly improving, particularly when it comes to simulating human-like reasoning.

In other words, just as blue-collar workers have seen their jobs threatened, we can expect those in better remunerated professions to start feeling the heat under their white collars as well. Whether the more powerful and privileged professional classes will defend their turf against AI any better remains to be seen. But one thing is clear – the bots are here to stay.

Charlotte Blease is the author of the forthcoming book, Dr Bot: Why Human Doctors Fail and How AI Can Save Lives (Yale University Press, 2025). Follow her on Twitter: @crblease