More GP training places have been created \u2013 the number is up 80 per cent in Ireland between 2019 and 2024 \u2013 but we still seem to be playing catch-up. Is there another possible solution?<\/p>\n
Northern Ireland researcher Charlotte Blease has been exploring the role artificial intelligence could play in future medical services. Author of one of this year\u2019s most mind-shifting reads, Dr Bot: Why Doctors Can Fail Us \u2013 and How AI Could Save Lives (Yale), Blease is not a medical doctor \u2013 her PhD is in philosophy \u2013 but she has worked extensively with medical professionals in areas from placebo studies to psychotherapy ethics.<\/p>\n
Facing facts that many of us would rather not think about it, she is this week\u2019s Unthinkable guest.<\/p>\n
Is AI now outperforming medical professionals in diagnosing or treating patients?<\/p>\n
\u201cIn some areas, yes \u2013 and faster than many clinicians realise. It depends on the specialty.<\/p>\n
\u201cIn radiology, for example, AI can already spot patterns and anomalies that the human eye simply can\u2019t detect. AI can read scans faster, more consistently, and in some cases more accurately than an average clinician \u2013 a game-changer for overstretched services and underserved regions. And while most AI systems aren\u2019t trained to \u2018think\u2019 like doctors, the evidence is piling up: GPT-4 \u2013 not even a medical model \u2013 can generate impressively accurate differential diagnoses, even in messy, complex cases. Medical-grade chatbots already can in some cases, and will, go further.<\/p>\n
\u201cWe\u2019re not at full replacement yet. But here\u2019s a hard home truth for healthcare: doctors make mistakes far more often than most people realise. Diagnostic errors occur in roughly 5 per cent to 20 per cent of clinical visits, and doctors only practice evidence-based medicine about half the time, even for common conditions. Mistakes disproportionately affect women, low-income patients, minorities and people with disabilities. This isn\u2019t about blaming clinicians \u2013 it\u2019s about recognising the limits of human cognition and the pressures of an overstretched system.\u201d<\/p>\n
Can you forecast a date when AI will be superior?<\/p>\n
\u201cI resist playing that game as there are too many variables . . . Technology improves. Humans, by contrast, haven\u2019t had a software update in 200,000 years. Our brains evolved for the `four Fs\u2019: feeding, fighting, fleeing and reproduction \u2013 not for parsing millions of data points or spotting microscopic signals in medical scans, nor for working in the frenetic, multitasking world of modern medicine. My message is: don\u2019t be dazzled by AI nor dismissive: it\u2019s important to recognise the direction of travel.\u201d<\/p>\n
![\"Philosopher](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2025\/12\/YZTUAOXSYJCMRIHRDXJTFEHGSA.jpg\")
Philosopher and researcher in medical ethics Charlotte Blease. Photograph: Jonas Fagerson People who are sceptical about AI displacing doctors talk of \u201cthe empathy problem\u201d. Is it overstated?<\/p>\n
\u201cDoctors like to think empathy is their superpower \u2013 but the evidence says otherwise. Social-science research shows clinicians routinely misread patients, and when you\u2019re burned out, rushed and juggling thousands of cases, empathy becomes performance. Most doctors are, frankly, the least well-equipped professionals to deliver it consistently. <\/p>\n
\u201cThe \u2018empathy problem\u2019 isn\u2019t overstated; it\u2019s misunderstood . . . AI won\u2019t replace human warmth, but it can help ensure patients feel heard, supported and not lost in the system.\u201d<\/p>\n
Is AI a solution to GP shortages?<\/p>\n
\u201cThe real question is always: compared to what? Not compared to instant access to Ireland\u2019s best GPs, because most people don\u2019t have that. With doctor shortages soaring and demand rising, the inverse care law is alive and well: those who need care most get it least. For these patients, it is AI compared to no help whatsoever. That\u2019s why patients are already turning to technology.\u201d <\/p>\n
Who will be held accountable if AI gives you bad advice or misdiagnoses you?<\/p>\n