Timeline of AI in medicine

From Timelines
Jump to navigation Jump to search

This is a timeline of AI in medicine.

Sample questions

The following are some interesting questions that can be answered by reading this timeline:

Big picture

Time period Development summary More details
1950s–1970s Early foundations AI in medicine begins with the development of rule-based expert systems that mimick clinical reasoning using logic and symbolic inference. In the 1950s and 1960s, pioneers like Alan Turing envision machines performing intelligent tasks. By the 1970s, Stanford’s MYCIN system demonstrates that computers can assist in diagnosing bacterial infections and recommending treatments, using a database of if-then rules. Though MYCIN never sees clinical use due to legal and trust concerns, it marks a foundational moment in AI-driven healthcare. The focus at this time is on replicating human decision-making through knowledge engineering, but systems struggle with ambiguity, learning, and real-world complexity.[1]
1980s–1990s AI Winter and Clinical Integration Interest in medical AI declines during the broader “AI Winter,” a period marked by limited computing power, high development costs, and unmet expectations. Early expert systems like INTERNIST-I and DXplain aim to support clinical diagnostics but face difficulties scaling beyond narrow domains. Researchers begin shifting from rigid logic systems to probabilistic models, such as Bayesian networks, which allow for better uncertainty management. Despite setbacks, AI continues to influence medical education and decision support in limited environments. This period focuses more on integration with hospital information systems than major technological breakthroughs, laying groundwork for more adaptable AI in future decades.
2000s–2010s Machine learning emerges The 2000s see a major shift from hand-coded rules to machine learning approaches, powered by growing clinical datasets and improved computing resources. With the spread of electronic health records and the internet, machine learning becomes viable for pattern recognition, predictive modeling, and risk assessment. Algorithms become increasingly used in radiology, oncology, and hospital operations. IBM Watson Health gains global attention for its potential in cancer diagnostics, although its real-world performance lags behind expectations. Despite some overhype, this decade marks a turning point where AI moves from theoretical promise to practical application, enabled by statistical models learning from large-scale clinical data.
2015–2020s Deep learning revolution With the rise of deep learning, especially convolutional neural networks (CNNs), AI made dramatic advances in fields requiring image analysis, such as dermatology, ophthalmology, and radiology. AI systems began matching or exceeding expert-level accuracy in diagnosing diseases from medical scans, such as diabetic retinopathy and skin cancer. This success was enabled by access to large labeled datasets and advances in GPU-based computing. Regulatory bodies like the FDA began approving AI-based diagnostic tools for clinical use. The era marked a growing acceptance of AI as a decision support tool, though interpretability and validation across populations remained key challenges.
2020s–present Generative AI and multimodal models The release of large language models (LLMs) like GPT-3 and GPT-4 spark a new phase in AI and medicine, emphasizing generative capabilities and human-like reasoning. These models are now used for summarizing patient records, assisting in clinical documentation, triaging, and even drafting research or guideline summaries. Meanwhile, multimodal AI systems—combining text, image, lab, and genetic data— enable more holistic, personalized diagnostics and treatment planning. Models like Med-PaLM and BioGPT demonstrate promising results in medical QA and education. However, challenges remain around transparency, clinical safety, bias, and regulatory frameworks. Still, AI becomes an integral assistant in healthcare workflows rather than just a predictive tool.

Full timeline

Year AI subfield Field of medicine Event type Event description
1959 Concept development Arthur Samuel coins the term "machine learning" while developing a self-learning checkers program.
1972 Expert systems and rule-based reasoning Infectious diseases (specifically blood infections / bacteremia and meningitis) and clinical decision support systems. Concept development Stanford University develops MYCIN, one of the earliest expert systems in artificial intelligence, designed to assist in diagnosing and treating blood infections. Using a knowledge base of approximately 500 “if-then” rules, MYCIN analyzes patient symptoms and test results, requests additional information if needed, and recommends treatments. It can also explain the reasoning behind its conclusions. MYCIN operates at a level comparable to medical specialists and outperforms general practitioners. As a pioneering rule-based system, it demonstrated the potential of AI in clinical decision-making and helped establish the foundation for future medical expert systems and AI applications in healthcare. An early expert system, MYCIN is foundational for medical decision support.[2]
1978 Causal reasoning and knowledge-based systems Ophthalmology (specifically glaucoma) and clinical decision support systems. Application The CASNET (causal-associational network) framework is introduced as one of the first AI systems based on causal models in medicine. Developed by Weiss, Kulikowski, Amarel, and Safir, CASNET models disease mechanisms through three components: patient observations, intermediate pathophysiological states, and disease classifications. Observations are linked to states, which form causal chains that map to specific diseases. These diagnostic conclusions then trigger general treatment recommendations, while detailed treatment strategies are tailored to the individual patient profile. The method is successfully applied to a consultation program for diagnosing and managing glaucoma, demonstrating CASNET’s value in clinical decision support.[3]
1984 Knowledge engineering and expert systems Internal medicine and diagnostics Concept development INTERNIST-1, one of the earliest comprehensive diagnostic expert systems, expands into the CADUCEUS project. Developed at the University of Pittsburgh, it uses structured medical knowledge to assist in diagnosing over 600 diseases, becoming a cornerstone in the development of AI-based diagnostic reasoning.
1992 (July) Probabilistic reasoning and decision-theoretic AI Pathology (specifically lymph-node diseases) and diagnostic decision support systems Milestone The Pathfinder Project emerges as an early example of AI assisting complex pathological decision-making. Developed by David Heckerman, Eric Horvitz, and Bharat Nathwani, Pathfinder is a normative expert system that aids in diagnosing lymph-node diseases using probability and decision theory. Initially, the team explores non-decision-theoretic approaches to handle uncertainty but finds them inadequate. Experimental and theoretical findings lead to a return to probabilistic methods, confirming their value for managing uncertain medical knowledge. Pathfinder demonstrates how decision-theoretic reasoning can enhance diagnostic accuracy and transparency in pathology, influencing the development of future expert systems in healthcare.[4]
1996 Neural networks and pattern recognition Cardiology and risk prediction Research finding Researchers apply backpropagation-based neural networks to predict sudden cardiac death from ECG signals. This early study highlights the promise of neural models for complex signal interpretation, outperforming linear statistical methods in risk stratification.
2003 Bayesian networks and clinical decision support Neonatology and sepsis diagnosis Research finding A Bayesian belief network is implemented in neonatal intensive care units to detect early signs of sepsis. The system integrates lab values, vital signs, and clinical context to predict sepsis hours before clinical recognition, improving outcomes in high-risk infants.
2006 Deep learning and unsupervised learning Medical imaging and computational medicine Research finding Geoffrey Hinton and his team introduce Deep Belief Networks (DBNs), marking a major breakthrough in the development of deep learning, and setting the stage for modern neural network applications in medical imaging. DBNs use an unsupervised layer-wise training method, making it possible to train deep neural networks more effectively than before. This approach enables the models to learn complex, high-level features from data, significantly improving tasks like object detection and speech recognition. The success of DBNs would transform deep neural networks from theoretical research topics into practical tools for solving real-world problems, laying the foundation for the rapid growth of modern deep learning applications across many fields.[5]
2010 Mobile health and AI triage Primary care and remote consultation Deployment Babylon Health launches in the UK, using rule-based triage algorithms to assess user symptoms via a mobile interface and recommend next steps. While controversial for accuracy, the service demonstrates AI’s potential in direct-to-patient triage and care guidance.
2011 Machine learning and diagnostic decision support Cardiology and emergency medicine Application Researchers at the Cleveland Clinic deploy a machine learning model to predict risk of heart failure and rehospitalization in patients using EHR data. The model identifies risk factors like lab values and vital signs, enabling earlier interventions and reduced readmission rates.
2012 Natural language processing and clinical decision support systems Oncology and clinical decision support Milestone Memorial Sloan Kettering Cancer Center (MSKCC) and IBM announce a collaboration to apply IBM's Watson technology to cancer care. The partnership aims to create an AI-powered clinical decision support tool that helps oncologists worldwide make personalized, evidence-based diagnostic and treatment decisions. By combining Watson’s natural language processing and rapid data analysis with MSKCC’s clinical expertise and vast cancer database, the system will deliver updated recommendations tailored to individual patients. This initiative addresses the growing complexity of cancer treatment and aims to accelerate the spread of cutting-edge oncology knowledge to improve outcomes across diverse healthcare settings.[6]
2014 AI in prosthetics and motor control Rehabilitation medicine and orthopedics Research milestone Researchers at Johns Hopkins University develop an advanced prosthetic arm controlled via neural signals and machine learning algorithms. The system uses pattern recognition to translate electrical activity from residual muscles into multi-degree movements, enhancing mobility and dexterity for amputees.
2014 (October) Computer vision and medical image analysis Radiology and diagnostic imaging Application AI startup Enlitic raises $2 million in seed funding to develop deep learning tools for medical diagnosis. Founded by Jeremy Howard, Enlitic aims to analyze large archives of medical images—like CT scans and X-rays—to assist doctors in identifying diseases more accurately and efficiently. Using artificial neural networks, the system learns patterns from existing medical data and generates predictions for new cases. With partnerships in the U.S., Brazil, China, and India, Enlitic also builds an "imaging analytics toolbox" to speed up algorithm development. The company sees particular potential in low-resource settings, where AI tools can greatly enhance diagnostic access.[7]
2015 Reinforcement learning and robotics Rehabilitation medicine Prototype demonstration Researchers at the Rehabilitation Institute of Chicago unveil a robotic exoskeleton that uses reinforcement learning algorithms to assist stroke patients in relearning motor skills. The exoskeleton adapts to each patient’s movements and provides real-time feedback, showing early promise in improving gait and muscle coordination. The project highlights the potential of AI-powered robotics in neurorehabilitation.
2015 AI for hospital logistics Clinical operations and patient flow Deployment Johns Hopkins Hospital launches the “Capacity Command Center,” an AI-powered operations hub that monitors patient flow, predicts bed availability, and optimizes staffing and discharge planning. The system integrates real-time data and machine learning to improve operational efficiency.
2016 NLP and clinical data mining Health records and medical informatics Research finding The i2b2 NLP challenge demonstrates that machine learning algorithms can extract medical problems, medications, and test results from unstructured clinical notes with near-human accuracy, paving the way for clinical AI systems that learn directly from existing patient data.
2016 Voice analysis and behavioral AI Psychiatry and neurodegenerative disease Research finding A study shows that machine learning algorithms can detect depression and early Parkinson’s disease from voice recordings with significant accuracy by analyzing vocal features such as pitch, tone, and pauses—offering promise for non-invasive mental health diagnostics.
2017 Deep learning and computer vision Dermatology and medical imaging Research finding Stanford researchers develop an AI system capable of detecting skin cancer with accuracy comparable to 21 expert dermatologists. Trained on a dataset of 129,450 images representing over 2,000 skin diseases, the deep learning algorithm successfully identifies various types of skin cancer, including melanoma and keratinocyte carcinoma. While at the time reliant on high-quality clinical images, the system holds potential for mobile deployment, expanding access to early diagnosis. Experts caution that more training with smartphone-quality images is needed. The study underscores AI's broader promise in medical imaging, with similar efforts underway in ophthalmology, oncology, and cardiovascular prediction.[8]
2017 AI and rare disease diagnosis Pediatrics and medical genetics Application Face2Gene, a facial analysis tool powered by deep learning, is adopted in genetics clinics to help identify rare syndromes based on facial morphology. The tool supports clinicians in recognizing over 300 genetic disorders and assists in shortening the diagnostic journey.
2018 Deep learning and medical image analysis Ophthalmology and medical imaging Milestone Researchers from DeepMind, University College London, and Moorfields Eye Hospital develop an AI system capable of detecting over 50 common eye diseases from 3D retinal scans with accuracy comparable to expert doctors. Using deep learning and optical coherence tomography (OCT) scans, the software can recommend patients for treatment based on its analysis. Although not yet approved for clinical use, it demonstrates its ability to assist in hospitals, helping prioritize urgent cases and improve diagnosis efficiency. The system’s potential is seen as transformative in managing sight-threatening conditions and can significantly expand access to early and accurate eye disease diagnosis globally.[9]
2018 AI and mobile retinal screening Ophthalmology and global health Field deployment IDx-DR, the first autonomous AI system for detecting diabetic retinopathy, is deployed in mobile screening units in rural India. The system enables early detection in underserved areas, where access to ophthalmologists is limited, and demonstrates AI’s potential in global health equity.
2018 Federated learning Radiology and medical imaging Concept introduction Google Research introduces the concept of federated learning in healthcare through a collaboration with clinics using Gboard. This privacy-preserving AI approach enables model training on decentralized medical data—such as imaging or EHRs—without moving sensitive patient records. The method addresses data silos and improves model generalizability while complying with privacy regulations.
2019 (February 1) Machine learning and medical data integration Pulmonology and radiology Data release MIT’s Laboratory for Computational Physiology releases the MIMIC-Chest X-Ray (MIMIC-CXR) database, offering over 350,000 anonymized chest radiographs collected from Beth Israel Deaconess Medical Center over five years. Free and open to academic, clinical, and industrial researchers via PhysioNet, this is the largest public repository of its kind. The database supports the development of AI models to detect conditions like pneumonia, cardiomegaly, and edema—especially beneficial for underserved areas lacking radiologists. The project aims to link imaging data with clinical records (via MIMIC-III), enabling more robust diagnostic tools. Collaboration with Stanford enhances the dataset’s generalizability across different healthcare contexts.[10]
2019 (February 14) Natural language processing and generative modeling Medical informatics and health communication LLM integration OpenAI's GPT‑2 is released as a large-scale unsupervised language model with 1.5 billion parameters, trained to predict the next word using 40GB of internet text. Without task-specific training, GPT‑2 achieves impressive results in language modeling, reading comprehension, summarization, translation, and question answering. It generates coherent, contextually adaptive text but still shows occasional errors. Due to concerns about potential misuse—such as disinformation, impersonation, and spam—OpenAI initially withheld the full model, opting for a staged release strategy. GPT‑2 highlights the potential and risks of large language models, prompting broader discussions on responsible publication, AI policy, and the societal implications of advanced text-generation systems.[11]
2019 Deep learning and rare disease discovery Neurology and genomics Application Researchers apply deep learning to analyze genomic sequences and uncover unknown variants associated with rare neurological disorders. The method accelerates the process of finding genotype-phenotype correlations, especially in pediatric neurology.
2019 Research finding MIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports[12]
2020 COVID-19 Response AI models used for COVID-19 spread prediction, CT scan interpretation, and drug discovery.[13]
2020 Explainable AI (XAI) Oncology and diagnostics Research finding A study published in *Nature Machine Intelligence* applies saliency maps and attention-based explainability tools to mammography AI systems. The research shows that transparent models improve trust and allow radiologists to better understand how AI reaches its conclusions, an essential step toward clinical adoption.
2020 Wearable AI and sensor fusion Preventive medicine and cardiovascular health Application Apple Watch Series 6 introduces blood oxygen monitoring supported by machine learning algorithms. Combined with ECG and heart rate variability data, the system enables early detection of atrial fibrillation and respiratory issues, offering continuous at-home monitoring and potential integration with physician workflows.
2020 Computational biology and structural bioinformatics Molecular biology and drug discovery Breakthrough AlphaFold revolutionizes biology by accurately predicting the 3D structures of proteins—life’s essential building blocks—in just minutes. Developed by DeepMind, AlphaFold solves a decades-old scientific challenge, dramatically reducing the time and cost needed to determine protein shapes, which previously had taken years and vast resources. By revealing millions of intricate protein structures, AlphaFold enables scientists to better understand how proteins function and interact with other molecules. This breakthrough accelerates research across medicine, biology, and environmental science, allowing scientists to focus their efforts on developing treatments and solutions to some of society’s most pressing challenges.[14]
2021 AI and mental health Psychiatry and digital therapeutics Application Wysa, an AI chatbot for mental health support, receives approval as a Class I medical device in the UK. The tool uses cognitive behavioral techniques, sentiment analysis, and NLP to help users manage anxiety and depression. It marks one of the first instances of a regulatory body approving an AI-driven mental health assistant.
2021 Swarm learning and privacy-preserving AI Oncology and decentralized data analysis Concept demonstration Researchers introduce “swarm learning,” a decentralized machine learning approach where cancer diagnosis models are collaboratively trained across hospitals without data sharing. The method shows high accuracy while preserving privacy, offering an alternative to centralized federated learning.
2021 AI and wearable biosensors Sports medicine and preventive cardiology Product integration Fitbit integrates AI-based algorithms that predict atrial fibrillation events by analyzing sleep and heart rhythm data. The system improves preventive care by alerting users and physicians ahead of major cardiac episodes.
2021 Robotics and surgical assistance Surgery and operating room automation Deployment CMR Surgical's Versius robot, supported by AI-assisted motion planning and instrument control, is adopted in UK hospitals for minimally invasive surgery. The system reduces surgeon fatigue and procedure time, representing a growing trend in AI-enhanced robotic surgery platforms.
2022 Medical natural language processing and instruction-tuned language models Medical informatics and clinical reasoning LLM integration Google Research introduces Med-PaLM, the first large language model (LLM) specifically designed for medical reasoning. Built on earlier models (PaLM and Flan-PaLM), it is tested using MultiMedQA, a benchmark that includes several medical question-answering datasets. Flan-PaLM performs well, scoring 67.6% on USMLE-style questions, but still shows issues with accuracy and safety. Med-PaLM is improved using a method called instruction prompt tuning, which helps it give more accurate and safer answers. In some areas, its performance is close to that of human doctors.[15]
2022 Multimodal diagnostic models Oncology and radiopathology Research demonstration Researchers combine pathology slides and radiology images into a unified AI diagnostic model for breast cancer. The multimodal system improves diagnostic accuracy by integrating imaging and histology, showing synergistic benefits of cross-domain AI.
2022 Computer vision and workflow optimization Emergency department operations Deployment AI startup Qventus partners with U.S. hospitals to deploy real-time decision support systems using vision and operations data to predict ED bottlenecks and optimize bed availability, contributing to reduced wait times and more efficient resource use.
2022 (October) AI in drug discovery Oncology and precision medicine Milestone Insilico Medicine announces the first AI-designed drug to enter Phase I clinical trials: a fibrosis treatment generated via deep generative models. This milestone marks a turning point in how AI accelerates the drug discovery pipeline.
2023 (February 1) Clinical decision support and diagnostic reasoning Diagnostic medicine and triage decision support A study evaluates GPT-3’s ability to diagnose and triage medical conditions using 48 validated case vignettes. GPT-3 correctly includes the diagnosis in its top three suggestions for 88% of cases, outperforming laypeople (54%) but trailing physicians (96%). For triage decisions, GPT-3’s accuracy (71%) was similar to laypeople (74%) but notably lower than physicians (91%). The model shows reasonably well-calibrated confidence levels, with Brier scores of 0.18 for diagnosis and 0.22 for triage. While GPT-3 performs well in diagnosis without specialized training, its triage capabilities remain limited, suggesting potential for clinical support with further refinement.
2023 (May 5) Generative AI and clinical natural language processing Health informatics and clinical documentation Deployment Epic and Microsoft announce the integration of GPT-4 into electronic health records (EHRs), marking a significant step in applying generative AI to healthcare. The collaboration introduces two key AI-enabled features. First, clinicians can use In Basket to generate draft responses to patient messages, improving communication efficiency. Second, Slicer Dicer, Epic’s data visualization tool, now uses AI to suggest relevant metrics based on user queries, streamlining data analysis. These tools aim to enhance provider productivity and patient engagement, representing a practical application of advanced AI within clinical workflows. This is the first large-scale deployment of LLMs in healthcare documentation.[16]
2023 (July) Genomic AI and rare disease diagnostics Genomics and pediatrics Clinical deployment NVIDIA and Oxford University Hospitals pilot an AI tool that analyzes whole-genome sequencing data to identify rare disease variants in pediatric patients. The system reduces diagnostic time from months to days, improving early treatment and family planning.
2023 Causal inference and counterfactual AI Health economics and intervention modeling Research milestone A study in *Nature Communications* applies causal inference AI models to estimate counterfactual outcomes of health interventions using EHR data. This enables policy planners to simulate the impact of treatment strategies and allocate resources more efficiently.
2023 AI-based virtual health assistants Geriatrics and chronic disease care Application Catalia Health launches Mabu, an AI-powered, voice-enabled robot that engages patients with chronic conditions at home. Mabu checks medication adherence, symptoms, and emotional well-being, transmitting data to care teams and improving patient engagement and remote care.
2023 Computer vision and autonomous diagnostic systems Ophthalmology and primary care Regulatory action Eyenuk receives FDA clearance to expand its EyeArt AI system, allowing its use with the Topcon NW400 retinal camera alongside previously approved Canon models. EyeArt is the first FDA-cleared AI system compatible with multiple camera brands for autonomous detection of diabetic retinopathy (DR). The update includes Real-Time Image Quality Feedback and an enhanced image quality module, achieving best-in-class gradability without dilation. Clinical trials show high accuracy: 94.4% sensitivity for mild DR and 96.8% for vision-threatening DR. With over 230,000 patients screened globally, the system aims to make AI-powered eye exams more accessible in primary care settings.[17]
2023 (September 13) Self-supervised learning and foundation models Ophthalmology and cardiovascular diagnostics A study introduces RETFound, a foundation AI model designed for generalizable disease detection from retinal images. Trained on 1.6 million unlabelled retinal images using self-supervised learning, RETFound can be adapted efficiently to various diagnostic tasks with minimal labeled data. The model significantly outperforms existing approaches in detecting eye diseases and predicting systemic conditions such as heart failure and myocardial infarction. RETFound demonstrates strong potential to improve diagnostic accuracy while reducing reliance on expert annotations, offering a scalable, label-efficient framework for broad clinical applications in ophthalmology and beyond.[18]
2023 (October) Augmented reality and AI surgery Surgical education and orthopedics Deployment Microsoft HoloLens is integrated with AI-powered surgical guidance in orthopedic procedures, providing real-time holographic overlays to assist in screw placement and bone alignment. Early trials report reduced operating time and improved accuracy.
2023 (December) Clinical foundation models and zero-shot reasoning Medical diagnostics and general practice Research milestone Researchers at Stanford and Johns Hopkins test Med-Gemini, a multimodal large language model capable of diagnosing conditions from images and text without task-specific tuning. The model shows promising zero-shot accuracy on benchmark datasets, signaling a shift toward general-purpose diagnostic AIs usable across specialties.
2024 (January) Embedded AI systems and diagnostic device integration Dermatology and primary care The FDA approves DermaSensor, the first AI-powered, noninvasive diagnostic tool for skin cancer detection at the point of care. This wireless handheld device uses spectroscopy and an FDA-cleared algorithm to analyze lesions for over 200 types of skin cancer, including melanoma, basal cell carcinoma, and squamous cell carcinoma. Clinical trials involving over 1,000 patients, led by the Mayo Clinic, demonstrated 96% sensitivity and 97% negative predictive value. A companion study with physicians showed DermaSensor halved missed cancer cases. This breakthrough highlights the integration of AI and spectroscopy in improving early cancer detection in primary care settings.[19]
2024 (March) Large language models and medical reasoning Medical education and clinical decision support A study evaluates whether large language models (LLMs), such as GPT-3.5 and Llama 2, can reason effectively about complex medical questions. Using benchmarks like MedQA-USMLE, MedMCQA, and PubMedQA, and testing methods including chain-of-thought prompting and retrieval augmentation, the authors found that GPT-3.5 reached passing scores on all three datasets. InstructGPT demonstrated the ability to recall and reason with expert medical knowledge. Open-source models like Llama 2 are also closing the performance gap. The findings suggest that with proper prompting, LLMs can support medical decision-making, though challenges such as uncertainty quantification and positional bias remain.[20]
2024 (April) AI-assisted home care monitoring Geriatrics and remote monitoring Pilot program Japan launches a national initiative deploying AI-enhanced ambient sensors in senior residences to detect falls, behavioral anomalies, and health deterioration. Machine learning algorithms analyze activity patterns to notify caregivers in real time.
2024 (May) AI for clinical workflow automation Hospital operations and patient safety Deployment Amazon Web Services launches AWS HealthScribe, a HIPAA-eligible service that uses speech recognition and NLP to generate structured clinical notes from patient-provider conversations. Integrated with EHR systems, it reduces clerical workload, minimizes transcription errors, and supports burnout reduction in clinical settings.
2024 (June) Generative AI and patient education Public health communication Deployment Mayo Clinic debuts an AI chatbot for patient education that explains test results, diagnoses, and procedures in plain language. The system, powered by a fine-tuned LLM, increases patient understanding and satisfaction scores while reducing call center volume.
2024 Real-time AI and surgical robotics Neurosurgery and intraoperative navigation Clinical application A real-time AI-enhanced surgical guidance system is tested in neurosurgical operations, using intraoperative imaging and predictive modeling to help surgeons avoid critical structures and improve outcomes. Early trials show improved precision and reduced complications.
2024 AI ethics and multimodal learning Global health and health policy Policy & ethics The World Health Organization (WHO) releases new guidance on the ethics and governance of large multi-modal models (LMMs)—AI systems that process diverse data types (text, images, video) and are increasingly used in health care. The report offers over 40 recommendations for governments, tech companies, and healthcare providers to ensure responsible use. Benefits of LMMs include support in diagnosis, patient communication, education, and research. However, WHO warns of risks like misinformation, bias, data quality issues, automation bias, and cybersecurity threats. WHO calls for inclusive development, regulatory oversight, post-deployment audits, and public infrastructure to promote ethical and equitable AI in health.[21]
2024 (August 1) Ethics & governance European Commission drafts AI Act with special provisions for medical AI systems.[22]
2024 (November) AI in pathology and rare cancers Oncology and histopathology Research breakthrough Researchers develop a foundation model for histopathology called PathGPT, trained on millions of whole-slide images. It achieves strong generalization across rare cancer subtypes and different institutions, enabling assistive diagnostics in under-resourced labs.
2025 (January 15) Multimodal learning Radiology and genomic medicine Mayo Clinic announces major partnerships with Microsoft Research and Cerebras Systems to advance AI in healthcare. With Microsoft, Mayo develops multimodal foundation models using chest X-rays and radiology reports to enhance diagnostics, automate workflows, and improve patient care. In parallel, Mayo and Cerebras create a genomic foundation model that uses exome and genome data to personalize treatments, with early success in predicting rheumatoid arthritis therapy responses. These collaborations leverage powerful AI and computing technologies to accelerate diagnosis, improve clinical precision, and bring personalized medicine closer to everyday care through scalable, real-world applications.[23]
2025 (February) AI benchmarking and risk assessment Health AI governance Policy & regulation The U.S. FDA launches the HealthAI Benchmark Suite (HABS), a standardized set of tasks and datasets to evaluate safety, fairness, and accuracy in AI tools for diagnostics, triage, and treatment recommendation. HABS serves as a precursor to adaptive regulation, allowing for pre- and post-market performance auditing of clinical AI tools.
2025 (March) AI for personalized nutrition Endocrinology and metabolic health Commercial deployment A startup-backed AI app integrates continuous glucose monitoring, dietary logs, and metabolic modeling to offer real-time personalized meal advice to diabetics and pre-diabetics. Clinical studies show improved glycemic control and patient adherence.
2025 (May) Large language models and differential diagnosis Emergency medicine and general practice Research finding A study finds that GPT-4, when prompted with structured clinical case formats, matches or exceeds junior doctors in generating differential diagnoses in emergency medicine cases. However, the model shows overconfidence on rare cases, reinforcing the need for physician oversight.
2025 (June) Digital twins and simulation-based medicine Cardiology and precision health Prototype deployment A hospital system in Germany launches a digital twin platform for cardiac patients, combining patient-specific imaging, genetics, and real-time sensor data to simulate disease progression and optimize personalized therapy decisions in silico.

Meta information on the timeline

How the timeline was built

The initial version of the timeline was written by Sebastian.

Funding information for this timeline is available.

Feedback and comments

Feedback for the timeline can be provided at the following places:

  • FIXME

What the timeline is still missing

  • Column for AI subfield
  • Field of medicine

Timeline update strategy

See also

References

  1. ↑ "Heuristic DENDRAL: A program for generating explanatory hypotheses in organic chemistry". ResearchGate. NASA Technical Reports Server (NTRS). February 1968. Retrieved 21 June 2025.
  2. ↑ Template:Cite encyclopedia
  3. ↑ Weiss, Sholom M.; Kulikowski, Casimir A.; Amarel, Saul; Safir, Aran (August 1978). "A model-based method for computer-aided medical decision-making". Artificial Intelligence. 11 (1–2): 145–172. doi:10.1016/0004-3702(78)90015-2. Retrieved 2025-06-07.
  4. ↑ Heckerman, David E.; Horvitz, Eric J.; Nathwani, Bruce N. (July 1992). "Toward Normative Expert Systems: Part I. The Pathfinder Project". Methods of Information in Medicine. 31 (2): 90–105. doi:10.1055/s-0038-1634867. Retrieved 2025-06-07.
  5. ↑ Ahmed, Sahin (9 October 2024). "Geoffrey Hinton: The Godfather of AI Who Now Warns of Its Dangers". Medium. Retrieved 2025-06-07.
  6. ↑ "Memorial Sloan Kettering Cancer Center, IBM to Collaborate in Applying Watson Technology to Help Oncologists". Memorial Sloan Kettering Cancer Center. 22 March 2012. Retrieved 2025-06-07.
  7. ↑ Novet, Jordan (28 October 2014). "Enlitic picks up $2 M to help diagnose diseases with deep learning". VentureBeat. Retrieved 2025-06-07. {{cite web}}: Check date values in: |date= (help); no-break space character in |date= at position 11 (help); no-break space character in |title= at position 17 (help)
  8. ↑ "This AI can spot skin cancer as accurately as a doctor". WIRED. 25 January 2017. Retrieved 2025-06-07.
  9. ↑ "DeepMind's AI can detect over 50 eye diseases as accurately as a doctor". The Verge. 13 August 2018. Retrieved 2025-06-07.
  10. ↑ Young, Annie (1 February 2019). "MIMIC Chest X‑Ray database to provide researchers access to over 350,000 patient radiographs". MIT News. Retrieved 2025-06-07. {{cite web}}: Check date values in: |date= (help); no-break space character in |date= at position 2 (help); no-break space character in |title= at position 12 (help)
  11. ↑ Radford, Alex, Jeffrey Wu, Dario Amodei, Daniela Amodei, Jack Clark, Miles Brundage, Ilya Sutskever (14 February 2019). "Better language models and their implications". OpenAI. Retrieved 2025-06-07.{{cite web}}: CS1 maint: multiple names: authors list (link)
  12. ↑ Johnson, Alistair E. W.; Pollard, Tom J.; Berkowitz, Seth J.; Greenbaum, Nathaniel R.; Lungren, Matthew P.; Deng, Chih‑ying; Mark, Roger G.; Horng, Steven (12 December 2019). "MIMIC‑CXR, a de‑identified publicly available database of chest radiographs with free‑text reports". Scientific Data. 6. doi:10.1038/s41597-019-0322-0. Retrieved 2025-06-07. {{cite journal}}: Check date values in: |date= (help); no-break space character in |date= at position 3 (help)
  13. ↑ "Predicting a pandemic: How AI helped predict COVID-19's twists and turns". VaccinesWork. Gavi, the Vaccine Alliance. 28 March 2022. Retrieved J2025-06-07. {{cite web}}: Check date values in: |access-date= (help)
  14. ↑ "AlphaFold". DeepMind. Retrieved 2025-06-07.
  15. ↑ "Med-PaLM: A Medical Large Language Model". Google Research. Retrieved 2025-06-07.
  16. ↑ "Epic and Microsoft Bring GPT-4 to EHRs". Epic. 5 May 2023. Retrieved 2025-06-07.
  17. ↑ "New FDA Clearance Makes Eyenuk the First Company with Multiple Cameras for Autonomous AI Detection of Diabetic Retinopathy". Eyenuk. 22 June 2023. Retrieved 2025-06-07.
  18. ↑ Zhou, Yukun; Chia, Mark A.; Wagner, Siegfried K.; Ayhan, Murat S.; Williamson, Dominic J.; Struyven, Robbert R.; Liu, Timing; Xu, Moucheng; Lozano, Mateo G.; Woodward‑Court, Peter; Kihara, Yuka (13 September 2023). "A foundation model for generalizable disease detection from retinal images". Nature. 622 (7981): 156–163. doi:10.1038/s41586-023-06555-x. Retrieved 2025-06-07. {{cite journal}}: Check date values in: |date= (help); no-break space character in |date= at position 3 (help)
  19. ↑ "FDA Approves First AI‑Powered Skin Cancer Diagnostic Tool". AIM at Melanoma Foundation. 18 January 2024. Retrieved 2025-06-07. {{cite web}}: Check date values in: |date= (help); no-break space character in |date= at position 3 (help)
  20. ↑ Liévin, Valentin; Hother, Christoffer Egeberg; Motzfeldt, Andreas Geert; Winther, Ole (1 March 2024). "Can large language models reason about medical questions?". Patterns. 5 (3): 100943. doi:10.1016/j.patter.2024.100943. Retrieved 2025-06-07. {{cite journal}}: Check date values in: |date= (help); no-break space character in |date= at position 2 (help)
  21. ↑ "WHO releases AI ethics and governance guidance for large multi-modal models". World Health Organization. 18 January 2024. Retrieved 2025-06-07.
  22. ↑ "AI Act enters into force". European Commission. 1 August 2024. Retrieved 2025-06-07.
  23. ↑ "Mayo Clinic partners with Microsoft Research and Cerebras to revolutionize AI in healthcare". News‑Medical. 15 January 2025. Retrieved 2025-06-07.
Retrieved from "http://timelines.issarice.com/index.php?title=Timeline_of_AI_in_medicine&oldid=85907"