Abstract
AI is transforming healthcare by reshaping diagnostics, streamlining healthcare operations, and offering new pathways to support elders. This article examines current and future applications of AI in healthcare, with a focus on older adults and the care continuum. It highlights Explore Digits’ approach to ethically leveraging AI to promote strength-based aging, enhance caregiver support, and drive equitable innovation. By centering older adults as co-creators in AI design and development, Explore Digits is advancing scalable solutions that elevate wellness, independence, and dignity. A hybrid future—where AI augments human care—offers the most promise for compassionate, inclusive healthcare.
Key Words
Artificial Intelligence, strength-based aging, geriatrics, healthcare transformation, human-centered AI, ethics, caregiver support, predictive healthcare
AI is reshaping healthcare at an unprecedented pace, offering solutions that enhance diagnostics, improve efficiency, and personalize treatment. As these innovations take hold, their impact on older adults—who often face complex health challenges—will be profound. Companies like Explore Digits are leading this transformation, leveraging AI to drive better care outcomes, support caregivers, and revolutionize how healthcare systems operate.
This article explores the current applications of AI in healthcare, forecasts the evolution of AI’s role in healthcare over the next decade and beyond, and highlights Explore Digits’ role in this paradigm shift.
The Current AI Landscape in Healthcare
AI in Diagnostics and Decision Support
Across the past decade, AI has steadily improved its ability to support diagnostic decision-making. In controlled research settings, algorithms have equaled or surpassed human experts in detecting specific conditions such as diabetic retinopathy and breast cancer (McKinney et al., 2020). A randomized controlled trial (RCT) found that AI breast cancer screening matched the standard double reading by two radiologists in detection rates and improved workflow efficiency (Lang, 2023). Newer models continue to push the envelope, especially in pattern-rich domains like neurology. For example, in dementia research the shift to complex neural networks has improved diagnostic accuracy for dementia detection compared to traditional methods and holds promise for earlier diagnoses and prognoses (Borchet et al., 2023; Lee et al., 2024).
Large language models (LLMs) like ChatGPT also have demonstrated promising—but mixed—results in clinical reasoning tasks. In a 2023 study comparing physician and AI chatbot responses to patient questions on Reddit, evaluators preferred the chatbot’s responses 79% of the time and rated them nearly 10 times more empathetic (Ayers et al., 2023). Meanwhile, Google’s Articulate Medical Intelligence Explorer outperformed primary care physicians in diagnostic accuracy and empathetic communication in 149 simulated text-based clinical scenarios (Tu et al., 2025).
These findings challenge the assertion that the medical establishment will be relatively undisrupted by AI due to distrust of AI by the public or inherent shortcomings of AI tools (Thirunavukarasu, 2023). Rather, the findings, while early, suggest that AI may not only provide accurate diagnoses but also may engage in meaningful, patient-centered communications.
Still, these models do not yet consistently perform. In real-world–inspired evaluations the most widely known LLM, ChatGPT, only correctly diagnosed 17 out of 100 cases from published pediatric case studies (Barile et al., 2024). Moreover, a 2020 systematic review found that only 5% of AI evaluations in healthcare used real patient care data (Bedi, 2025). This underscores the need for rigorous validation in diverse clinical settings.
Encouragingly, the evidence base is shifting from retrospective model assessments to prospective trials. A 2024 Lancet Digital Health review found that 81% of AI-related RCTs reported improved diagnostic yield or workflow efficiency. Most of these studies involved AI used alongside clinicians—not replacing them—highlighting a hybrid model as the likely path forward (Han, 2024).
AI in Clinical Care and Healthcare Operations
While diagnostic accuracy has garnered most of the media attention, we believe that greater opportunity lies in how AI can reshape the infrastructure of care for older adults—not only improving detection of disease but proactively enhancing wellness, equity, and autonomy.
Also, AI streamlines administrative tasks such as appointment scheduling, billing, and medical documentation. Healthcare chatbots are already widely used by providers, payers, and other groups. These tools provide real-time information access to patients and caregivers. More fundamental shifts in AI use in healthcare operations are being tested and implemented. The Mayo Clinic recently expanded its use of AI-assisted medical documentation wherein an AI tool uses recorded provider–patient interactions and drafts clinical notes (Olsen, 2025). Tampa General Hospital has deployed AI tools for patient monitoring since 2022, which has reportedly driven efficiencies like decreased waiting times for imaging and saved hundreds of lives by preventing sepsis (Glenfield, 2025). Continued research and development efforts in these areas have the potential to transform the healthcare experience, reducing burden and strain on the provider and patient alike and providing more space for human connection that is critical in a clinician/patient relationship.
Stanford University researchers are using AI-driven remote monitoring in diabetic patients to provide more robust indicators of metabolic health (Armitage, 2025; Metwally, 2024). The abilities of AI-powered cancer detection tools have greatly expanded in recent years. New models developed by researchers at Harvard are no longer limited to specific tasks or cancer types; rather, new models, trained on images from 19 different cancer sites, can not only detect cancer with improved precision compared to prior models, but also can identify individualized treatment options based upon the tumor’s cellular environment. (Wang et al., 2024). AI’s predictions of atrial fibrillation and stroke have also improved dramatically, offering society better tools to detect and prevent these conditions (Ding, 2020; Hampton, 2021).
‘We believe that greater opportunity lies in how
AI can reshape the infrastructure of care
for older adults … proactively
enhancing wellness, equity, and autonomy.’
In conclusion, the medical and social science literature portrays a rapidly evolving field: Early exuberance has been tempered by rigorous trials, which in turn are guiding more refined, powerful AI applications. If current trends continue, we can expect AI’s accuracy and practicality to further improve and that its role in clinical decision-making and healthcare operations will expand—likely not as a replacement for human clinicians, but as a valuable ally to enhance patient care (Shiwani, 2023). The best outcomes for older adults and patients will likely emerge from this synergy, where human empathy and understanding meet machine precision and data-driven insights.
Building upon this foundation, the next section explores how AI’s impact may unfold over time. We acknowledge that such predictions are fraught, but we also believe these conversations are necessary to shape the future we hope for and to avoid the many potential pitfalls along the way.
The Future of AI in Healthcare: A Time Horizon Perspective
Looking ahead, a hybrid approach—where AI helps human clinicians rather than replacing them—may offer the most viable path forward. While AI has demonstrated impressive diagnostic capabilities and empathetic communication, humans in the loop remain crucial to ensure AI-driven healthcare aligns with patient needs, ethical considerations, and clinical realities. Ongoing regulatory oversight, rigorous real-world evaluations, and structured physician-AI collaboration will be essential to achieving safe and effective AI integration into medical practice.
Looking forward, what can we expect? AI will continue to get “smarter” as more data (including real-world data from electronic health records and wearable devices) are used to train models. This could be particularly helpful in complex diagnostic scenarios common in older adults, such as multi-morbidity (where an AI can help parse which of several coexisting conditions is causing current symptoms).
Continuous learning systems might adapt over time to new data, maintaining or improving accuracy as disease patterns change or new treatments alter presentations. Moreover, explainable AI techniques are being developed so that models can highlight the reasons for a prediction (for example, marking the specific area on an X-ray that led to a diagnosis of pneumonia)—this could increase clinician trust and enable safer use of AI’s recommendations.
Despite these promises, the future also will require concerted efforts to address known pitfalls of AI technology. Efforts are underway to improve the reporting and validation standards for AI in healthcare, ensuring that claims of accuracy are backed by solid evidence and that models undergo independent evaluation on external cohorts. Regulators like the FDA have issued guidelines for algorithmic bias and are encouraging prospective trials for AI-based devices.
Experts also recognize that AI systems need monitoring post-deployment, just like drugs after approval, to catch any degradation in performance or unsafe failure modes, especially as they encounter cases with complexities beyond their original training. Of course, there is always the possibility of negative consequences of AI that can occur as an unintended consequence of systems or by design. A recent high-profile example where a large healthcare insurer allegedly used an AI system with known technical issues to determine medical necessity for healthcare services highlights the potential risks of AI in healthcare (Napolitano, 2023).
The field of AI in healthcare is evolving rapidly, making long-term predictions inherently uncertain. However, by examining current trends, ongoing research, and expert insights, we can make informed projections about how AI will shape healthcare in the coming years. Below, we outline some of the most promising advancements expected over different time horizons, recognizing that unexpected breakthroughs or regulatory changes could accelerate or slow adoption.
1–2 Years: The Immediate Future of AI in Healthcare (2025–2027)
Clinical Integration and Decision Support
AI-based clinical decision support systems will become more deeply integrated into electronic health records, offering real-time recommendations to physicians. Explore Digits is developing ML models to assist regulatory agencies in recognizing patterns in nursing home compliance and care patterns using unstructured text data collected during site inspections. This highlights the power of natural language processing (NLP) that can be used to extract insights from the rich, but underused text data that is ubiquitous in healthcare.
Expansion of AI in Drug Discovery
AI-driven simulations reduce the time required for preclinical trials, identifying promising compounds faster than traditional methods had. Companies are already leveraging AI-driven drug design, with the first AI-created drug molecules entering human trials.
AI-Driven Care of Older Adults
AI-powered virtual assistants will become essential companions for older adults, reminding them to take medication, scheduling doctor visits, and offering cognitive engagement tools for dementia care. Explore Digits is working on AI-driven support tools to help all stakeholders in healthcare, from policymakers to caregivers. In the near future, we envision that older adults will have virtual geriatricians on call in their smartphones to provide person-centered clinical recommendations.
2–5 Years: The Medium-Term Horizon (2027–2030)
Predictive Medicine and AI-Driven Prevention
AI’s ability to analyze longitudinal health data will make predictive medicine a reality. AI models will predict an individual’s likelihood of developing chronic conditions, allowing for early intervention. Explore Digits is investing in AI models that proactively identify risk factors for negative outcomes, including using ML to identify when an older adult is becoming frailer, potentially before caregivers or even family members would notice.
Robotics and AI in Surgery and Recovery
AI-assisted robotic surgery will become more autonomous, allowing for ultra-precise procedures with minimal human oversight. AI-powered rehabilitation robots will aid in physical therapy, helping older adults regain mobility post-surgery or after strokes. This promises to alleviate current staffing shortages in the healthcare system and provide more direct care to older adults at a fraction of the cost.
‘AI-assisted robotic surgery will become more autonomous, allowing for ultra-precise procedures with minimal human oversight.’
AI in Mental Health and Loneliness Prevention
AI-driven mental health chatbots and virtual therapists will provide accessible support for older populations, reducing the burden on human providers. AI also will play a role in social engagement by recommending activities, connecting older adults to communities, and identifying signs of depression or cognitive decline. To preserve human empathy amid AI advancement, developing compassionate AI will be a critical focus.
5–10 Years: Long-Term AI Impact (2030–2035)
AI as a Primary Healthcare Provider
By 2035, AI-driven primary care models could handle routine diagnoses and prescriptions, allowing human doctors to focus on complex cases. AI will provide instant, high-quality medical consultations, ensuring even the most chronically underserved populations have access to high quality healthcare. Cloud-based systems should increase access to specialists and all clinicians by providing virtual visits and support that mimic in-person interactions while decreasing the burden on patients and providers. AI should be integrated throughout the care continuum to allow more support to be delivered in people’s homes and communities, overcoming the longstanding institutional bias that previously made long-term care placement common.
Hospital-at-Home and AI-Based Care Coordination
The hospital-at-home model—where AI coordinates remote care teams, wearable devices, and virtual doctors—will become widespread. AI will guide patients and caregivers through complex medical journeys, predicting health risks and proactively deploying resources.
Advanced AI-Driven Robotics
AI-driven companion robots will offer emotional support, assist with daily activities, and provide mobility support for older adults (e.g., preventing falls or restoring the ability to ambulate), extending older adult independence while reducing caregiver burden. Importantly, these robots should be designed with strength-based approaches to support older adults to accommodate personalized goals.
Beyond 10 Years: The AI-Integrated Healthcare Ecosystem (Post–2035)
Fully Autonomous AI Healthcare Systems
By the late 2030s, AI will be deeply embedded in every aspect of healthcare, from personalized prevention plans to autonomous diagnostic centers. AI-driven global health monitoring will predict disease outbreaks in real-time, allowing for faster pandemic response and resource deployment.
Ethical and Regulatory Challenges
Despite AI advancements and their promise for a better future, concerns around bias, data security, equitable distribution, and ethical AI deployment will remain paramount. AI models must be continuously audited for fairness, particularly for older adults who may be disproportionately affected by AI-driven decision-making and bad actors.
One of the most pressing issues is algorithmic bias, which can arise from biased training data or flawed model design. Older adults, particularly those from underrepresented backgrounds, may face disparities in AI-generated recommendations for treatments, diagnoses, or care plans. To address this, regulatory bodies must enforce transparency in AI model development, requiring developers to disclose training data sources, demographic representation, and bias mitigation strategies.
Another significant challenge is data security and patient privacy. With AI processing vast amounts of sensitive health data, there is an increased risk of breaches or unauthorized data use. Healthcare institutions and AI developers must implement robust cybersecurity protocols, including encryption, decentralized data storage, and stringent access controls. Legislative measures, such as strengthening HIPAA protections, will be critical in ensuring patient trust in AI systems.
Furthermore, explainability and accountability in AI decision-making remain ongoing concerns. Many AI-driven diagnostic or decision-support systems function as “black boxes,” making it difficult for healthcare professionals to understand how AI arrived at a conclusion. To enhance trust and usability, AI developers should prioritize interpretable AI models that provide clear rationales for their outputs. Clinicians must be trained to critically assess AI-generated recommendations rather than relying on them unconditionally. Explainable AI has gained traction in recent years to uncover the inner workings of deep learning models. Human-annotated datasets are critical to ensure that models respond appropriately, however they often lack the necessary diversity (factual and counter-factual statements) to truly test how well models behave instead of relying on synthetically generated examples (Wu, 2021).
‘Ongoing human oversight and intervention must
be a cornerstone of AI-driven healthcare.’
To ensure ethical AI deployment in healthcare, a multi-stakeholder approach is necessary. Policymakers, healthcare providers, AI researchers, and patient advocacy groups must collaborate to create clear guidelines for AI accountability, develop standardized performance benchmarks, and establish regulatory sandboxes where AI innovations can be tested in controlled environments before widespread implementation.
Lack of access and health inequity may be alleviated by AI if we believe and insist on a future where AI creates more abundance, but there is also the risk that AI and AI-driven technologies will become commodities that only some can afford, which could lead to more stark health disparities.
Finally, ongoing human oversight and intervention must be a cornerstone of AI-driven healthcare. AI should augment, not replace, human decision-making, ensuring that healthcare remains patient-centered, equitable, and aligned with ethical medical practices.
Bringing the Vision to Life: Explore Digits’ AI Focus Areas
Explore Digits was founded in 2019 with a mission to harness the transformative power of AI and ML to radically improve the lives of older adults. The company was born from personal experiences with the healthcare system’s failure to support older adults with dignity near the end of life—combined with a track record of using cutting-edge technologies to solve complex, system-level problems. Long before the mainstream AI wave, we recognized that these tools could be used not just to treat illness, but to elevate wellness, independence, and dignity throughout the aging journey.
Our belief that AI can radically transform healthcare into a system that acknowledges and builds upon older adults’ strengths while limiting declines has only intensified with rapid AI developments and increased attention. Explore Digits aims to serve as a bridge between modern AI tools and real-world aging challenges—focusing on actionable, ethical, and scalable solutions for older adults, caregivers, and healthcare providers. Thus, we were honored to be invited to discuss this topic in Generations Journal.
At Explore Digits, we believe the future of AI in aging isn’t about replacing human care—it’s about reimagining care to be more personalized, empowering, and inclusive. AI can help amplify the voices of older adults, convert complex data into tailored support plans, and provide caregivers with timely, expert-level insights. But achieving this vision requires centering older adults not as passive recipients of innovation, but as co-creators.
Their lived experiences must inform every stage of the AI development process—from shaping datasets and training algorithms to refining interfaces and evaluating impact. When older adults guide the design, the result is technology that is not only smarter but also more dignified and humane.
We’re advancing this work through several core commitments:
- Promoting strength-based aging: AI should not only flag risks—it should also identify and reinforce resilience, capacity, and wellness.
- Enabling a “geriatrician-in-your-pocket”: AI-powered decision support can extend the reach of clinical expertise to caregivers and older adults, especially in underserved settings.
- Embedding older adults in the AI lifecycle: We champion their leadership in shaping data, testing applications, and informing ethical AI design.
Explore Digits was founded on the conviction that AI can be a transformative force for good in aging. We work alongside federal, state, and community partners, including the Centers for Medicare & Medicaid Services (CMS), nursing home providers, medical record review teams, and policy innovators—to build equitable and scalable solutions. Our tools are designed not only to interpret large volumes of healthcare and systems data but to highlight actionable insights that support strength-based care, proactive planning, and transparency across the continuum.
The table below highlights key opportunity areas where Explore Digits is applying AI to this mission, illustrating how we pair technical innovation with on-the-ground collaboration to improve the lives of older adults, caregivers, and care teams.
| Opportunity Area | Explore Digits Advantage | Examples and Work in Progress | Strategic Next Step |
| Gather insights from relatively underused data sources such as text, audio, and video to improve care and well-being. | Robust NLP and analytics engine capable of processing massive troves of data including unstructured text, audio, and video, and generating intelligent responses and recommendations. | Applications codeveloped with the federal government to analyze public sentiment on policies as well as analysis of regulatory compliance data collected by nursing home inspectors in the states. | Add AI-powered chatbot assistant for older adults, caregivers, and loved ones. This provides users with a virtual, expert-level, geriatric-trained provider and navigator that is “on call.” |
| Caregiver support and coordination | Data synthesis and visualization | Dashboards to identify trends in nursing home compliance and care quality | Expand to passive sensing data (e.g., data from wearable or voice monitoring that doesn’t require user support) to identify early signs of cognitive decline and well-being changes. |
| Chronic condition risk prediction | AI/ML applied to CMS and other healthcare datasets, improper payment evaluation | Using routinely collected data to indicate resiliency scores for older adults and predict the risk of adverse outcomes. | Co-develop personalized care applications to support older adults. Identify patterns in care to predict and prevent healthcare issues and negative outcomes. |
| Older adult–focused survey insights | Survey platform and NLP platform | Data collection system for nonprofit nursing homes to determine clinical outcomes associated with personalized care and culture change. | Launch sentiment tracker tied to health outcomes. |
| Financial/legal protections and support navigating complex social/medical systems | Secure, compliant IT infrastructure | Early stage fraud detection models | Partner with legal aid organizations and healthcare navigators for AI co-design. |
Conclusion: A Future Defined by AI and Human Collaboration
AI is set to redefine healthcare, offering remarkable opportunities and complex challenges. Companies like Explore Digits are ensuring that AI’s potential is harnessed responsibly, prioritizing ethical AI use, while delivering tangible benefits to people and providers alike to improve the overall healthcare ecosystem. By embracing AI’s promise—while remaining vigilant about its risks—we can create a smarter, more efficient, and more compassionate healthcare system that serves all generations.
Dan Andersen, PhD, MS, MPH, is Chief Health Informatics Officer; Faizan Wajid, PhD, is senior Data Scientist; and Ravi Hubbly is founder and CEO, all at Explore Digits in Rockville, MD.
Photo credit: Shutterstock/Nan_Got
References
Armitage, H. (2025). Researchers use AI to help predict and identify subtypes of Type 2 diabetes from simple glucose monitor. Stanford Medicine. https://med.stanford.edu/news/all-news/2025/01/type-2-diabetes.html
Ayers, J. W., Poliak, A., Dredze, M., Leas, E. C., Zhu, Z., Kelley, J. B. Faix, D. J., Goodman, A. M., Longhurst, C. A., Hogarth, M., & Smith, D. M. (2023). Comparing physician and Artificial Intelligence chatbot responses to patient questions posted to a public social media forum. JAMA Internal Medicine, 183(6), 589–596.
Barile, J., Margolis, A., Cason, G., Kim, R., Kalash, S., Tchaconas, A., & Milanaik, R. (2024). Diagnostic accuracy of a large language model in pediatric case studies. JAMA Pediatrics, 178(3), 313–315.
Bedi, S., Liu, Y., Orr-Ewing, L., Dash, D., Koyejo, S., Callahan, A., Fries, J. A., Wornow, M., Swaminathan, A., Lehmann, L. S., Hong, H. J., Kasheyap, M., Chaurasia, A., R., Shah, N. R., Singh, K., Tazbaz, T., Milstein, A., Pfeffer, M. A., & Shah, N. H. (2025). Testing and evaluation of health care applications of large language models: A systematic review. JAMA, 333(4), 319–328.
Borchert, R. J., Azevedo, T., Badhwar, A., Bernal, J., Betts, M., Bruffaerts, R., Burkhart, M. C., Dewarchter, I., Gellersen, H. M., Low, A., Lourida, I., Machado, L., Madan, C. R., Malpetti, M., Mejia, J., Michopoulou, S., Mūnoz-Neira, C., Pepys, J., Peres, M., Phillips, V., Ramanan, S., Tamburin, S., Tantiangco, H. M., Thakur, L., Tomassini, A., Vipin, A., Tang, E., Newby, D., Deep Dementia Phenotyping (DEMON) Network, Ranson, J. M., Llewellyn, D. J., Veldsman, M., & Rittman, T. (2023). Artificial intelligence for diagnostic and prognostic neuroimaging in dementia: A systematic review. Alzheimer’s & Dementia, 19(12), 5885–5904.
Ding, L., Liu, C., Li, Z. & Wang, Y. (2020). Incorporating Artificial Intelligence Into Stroke Care and Research. Stroke 51(12), e351–e354.
Glenfield, J. (2025). Hospitals across the country are using artificial intelligence. Scripps News. https://www.scrippsnews.com/science-and-tech/artificial-intelligence/hospitals-across-the-country-are-using-artificial-intelligence
Han, R., Acosta, J. N., Shakeri, Z., Ioannidis, J. P. A., Topol, E. J., Rajpurkar P. (2024). Randomised controlled trials evaluating artificial intelligence in clinical practice: A scoping review. The Lancet Digital Health, 6(5), e367–e373.
Hampton, T. (2021, November 15). Using AI to prevent blood clots, strokes. The Harvard Gazette. https://news.harvard.edu/gazette/story/2021/11/ai-based-method-predicts-risk-of-atrial-fibrillation/
Lang, K., Josefsson, V., Larsson, A.-M., Larsson, S., Högberg, C., Sartor, H., Hofvind, S., Andersson, I., & Rosso, A. (2023). Artificial intelligence-supported screen reading versus standard double reading in the Mammography Screening with Artificial Intelligence trial (MASAI): A clinical safety analysis of a randomised, controlled, non-inferiority, single-blinded, screening accuracy study. The Lancet Oncology, 24(8), 936–944.
Lee, Y. L., Vaghari, D., Burkhart, M. C., Tino, P., Montagnese, M., Li, Z., Zühlsdorff, K., Giorgio, J., Williams, G., Chong, E., Chen, C., Underwood, B. R., Rittman, T., & Kourtzi, Z. (2024). Robust and interpretable AI-guided marker for early dementia prediction in real-world clinical settings. eClinicalMedicine, 74, 102725. https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(24)00304-3/fulltext
McKinney, S. M., Sieniek, M., Godbole, V. Godwin, J., Antropova, N., Ashrafian, H., Back, T., Chesus, M., Corrado, G. S., Darzi, A., Etemadi, M., Garcia-Vincente, F., Halling-Brown, M., Hassabis, D., Jansen, S., Karthikesalingam, C., Kelly, C. J., King, D., Lesam, J. R., Melnick, D., Mostofi, H., Peng, L., Jay Reicher, J., Romera-Paredes, B., Sidbottom, R., Suleyman, M., Tse, D, Young, K. C., DeFauw, J., & Shetty, S. (2020). International evaluation of an AI system for breast cancer screening. Nature, 577, 89–94. https://www.nature.com/articles/s41586-019-1799-6
Metwally, A. A., Perelman, D., Park, H., Wu, Y., Jha, A., Sharp, S., Celli, A., Ayhan, E., Abbasi, F., Gloyn, A. L., McLaughlin, T., & Snyder, M. P. (2024). Prediction of metabolic subphenotypes of type 2 diabetes via continuous glucose monitoring and machine learning. Nature Biomedical Engineering. https://doi.org/10.1038/s41551-024-01311-6
Napolitano, E. (2023). UnitedHealth uses faulty AI to deny elderly patients medically necessary coverage, lawsuit claims. CBS News Money Watch. https://www.cbsnews.com/news/unitedhealth-lawsuit-ai-deny-claims-medicare-advantage-health-insurance-denials/
Olson, E. (2025). Mayo Clinic expands AI documentation partnership with Abridge. Healthcaredive. https://www.healthcaredive.com/news/abridge-mayo-clinic-ai-documentation-partnership/737339/
Shiwani, T., Relton, S., Evans, R., Kale, A., Heaven, A., Clegg, A., & Todd O. (2023). New horizons in artificial intelligence in the healthcare of older people. Age and Ageing, 52(12). https://academic.oup.com/ageing/article/52/12/afad219/7479755
Thirunavukarasu, A. J. (2023). Large language models will not replace healthcare professionals: Curbing popular fears and hype. Journal of the Royal Society of Medicine, 116(5), 181–182.
Tu, T., Schaekermann, M., Palepu, A., Saab, K., Freyberg, J., Tanno, R., Wang, A., Le, B., Amin, M., Cheng, Y., Vedadi, E., Tomasev, N., Azizi, S., Singhal, K., Hou, L., Webson, A., Kulkarni, K., Mahdavi, S. S. Semturs, C., Gottweis, J., Barral, J., Chou, K., Corrado, G. S., Matias, Y., Kathikesalingam, A., & Natarahan, V. (2025). Towards conversational diagnostic artificial intelligence. Nature. https://doi.org/10.1038/s41586-025-08866-7
Wang, X., Zhao, J., Marostica, E., Yuan, W., Jin, J., Zhang, J., Li, R., Tang, H., Wang, K., Li, Y., Wang, F., Peng, Y., Zhu, J., Zhang, J., Jackson, C. R., Zhang, J., Dillon, D., Lin, N. U., Sholl, L., Denize, T., Meredith, D., Ligon, K. L., Signoretti, S., Ogino, S., Golden, J. A., Nasrallah, M. P., Han, X., Yang, S., & Yu, K.-H. (2024). A pathology foundation model for cancer diagnosis and prognosis prediction. Nature 634, 970–978. https://www.nature.com/articles/s41586-024-07894-z
Wu, T., Ribeiro, M., Heer, J., Weld, D. (2021). Polyjuice: Generating counterfactuals for explaining, evaluating, and improving models. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Vol. 1; pp. 6707–6723).
