As healthcare costs rise and populations age, AI offers the promise of more efficient, personalised, and accessible care. Irish Medical Times recently published an article outlining 10 ways artificial intelligence will change medicine (Cosgrave, 2025).
1. Medical Imaging & Diagnostic Support
AI algorithms now assist radiologists by identifying regions of interest in X-rays, CT scans, and MRIs with high accuracy and speed. The AIdoc platform, used at Dublin’s Mater Misericordiae University Hospital, scan radiology studies in real time, flagging haemorrhages and pulmonary emboli. Google DeepMind and Zebra Medical Vision have developed eye disease detection AI algorithms that performed as well as expert ophthalmologists in Moorfields Eye Hospital in London.
2. Predictive Analytics for Patient Deterioration
Machine learning models have been used to analyse vital signs and Electronic Health Record (EHR) data to detect warning signs of sepsis, ICU readmission, or cardiac arrest. Kaiser Permanente’s neonatal sepsis calculator has reduced unnecessary antibiotic usage (Goel et al., 2020). Patients are 20% less likely to die of sepsis because of a new AI system developed at Johns Hopkins University. Using medical records and clinical notes, the AI system significantly cut patient mortality from one of the top causes of hospital deaths worldwide. The study is published in Nature Medicine (Saria, 2018).
3. Personalised Treatment Planning
AI-driven precision medicine uses genomic sequencing, biomarker profiles and clinical histories to simulate how a patient might respond to specific treatment. Companies like Tempus and IBM Watson for Oncology offer personalised cancer therapies with high concordance to expert recommendations. Personalised treatment plans can enable quicker initiation of therapy and improved treatment success rates. Companies such as Foundation Medicine and Guardant Health also use AI models to predict drug efficacy and optimal dosing regimens.
4. Virtual Health Assistants & Chatbots
Natural Language Processing (NLP) powered chatbots can help with symptom triage and medication reminders. This eases the workload of on call centres and clinics. Companies such as Babylon Health provide AI to assess symptoms and provide triage recommendations. The AI chatbot has matched or exceeded the diagnostic accuracy of seven experienced physicians. In the future this could potentially cut unnecessary clinic visits and call centre costs by up to 40%.
5. Remote Patient Monitoring & Wearables
Wearable devices like the Apple Watch and Dexcom glucose monitors use AI to provide real time healthcare monitoring. Apple watch’s ECG app can detect atrial fibrillation with sensitivity comparable to clinical ECG devices. Dexcom and Medtronic continuous glucose monitors use time-series AI models to predict hypoglycemic episodes, alerting patients up to 20 minutes before blood sugar drops to dangerous levels. In the UK, the NHS is piloting a sensor patch program that demonstrated a 30% reduction in readmissions over six months.
6. Accelerating Drug Discovery
AI models are used to predict molecular interactions, toxicity profiles and formulation stability, dramatically shortening drug development timelines. Insilico Medicine’s AI-designed molecule (INS018_055) for idiopathic pulmonary fibrosis is already in Phase II trials. This cuts candidate selection time from years to months, reduces R&D costs per candidate by up to 30%.
7. Administrative Workflow Automation
Robotic process automation (RPA) and NLP tools have the potential to change the future of administration workflow automation. Taken together, these can handle billing, coding, appointment scheduling and prior authorisation. It has been estimated that they can cut administrative workload by up to 50%. Olive AI and Nuance’s Dragon Medical Advisor are helping hospitals save millions, including by reducing denied claims by over 20%.
8. AI-Assisted Surgery & Robotics
Intuitive Surgical’s da Vinci 5 system and Medtronic’s Hugo robotic system use AI to refine instrument positioning, analyse tissue deformation and optimise suture placement. Computer vision AI enabled surgical robots to guide tool trajectories for minimally invasive procedures. Clinical trials conducted at MD Anderson and Mayo Clinic show a reduction of 15% in intraoperative complications, and shortened patient recovery times by an average of 2 days.
9. Clinical Documentation & Coding Support
Generative AI tools can also transcribe and summarise appointment interactions and suggest billing codes. Nuance’s Dragon Medical One has been integrated into over 77% of US hospitals, reducing documentation time by 60%. Microsoft’s Dragon Copilot is used in Mass General Brigham and Stanford Health. It uses audio capture to take notes and generate referral letters.
10. Guideline & Literature Query Tools
AI assistants like IBM Watson and UpToDate Edge retrieve clinical guidelines and literature in response to natural-language queries. This enables faster access for clinicians to provide evidence-based care and improved decision-making.
Conclusion
AI is reshaping every aspect of medicine, from diagnostics and treatment to administration and research. While its full potential is still unfolding, its ability to augment human expertise and free clinicians is already evident. The future of medicine will be deeply intertwined with AI, enabling better outcomes and more efficient care.
Reference
- Cosgrave, T. (2025). 10 ways that AI will change medicine. Irish Medical Times.
- Goel, N., Shrestha, S., Smith, R., Mehta, A., Ketty, M., Muxworthy, H., Abelian, A., Kirupaalar, V., Saeed, S., Jain, S., Asokkumar, A., Natti, M., Barnard, I., Pitchaikani, P. K., & Banerjee, S. (2020). Screening for early onset neonatal sepsis: NICE guidance-based practice versus projected application of the Kaiser Permanente sepsis risk calculator in the UK population. Archives of Disease in Childhood – Fetal and Neonatal Edition, 105(2), 118–122. https://doi.org/10.1136/archdischild-2018-316777
- Saria, S. (2018). Individualized sepsis treatment using reinforcement learning. Nature Medicine, 24(11), 1641–1642. https://doi.org/10.1038/s41591-018-0253-x
- Picture: A.BourgeoisP, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons
