The authors extend an established framework—Descriptive, Diagnostic, Predictive, and Prescriptive—with two new domains: Critique, and Creative. The result lays out where medical AI’s greatest opportunities lie, and how it could shift healthcare from treating disease to proactively reducing disease burden. At the same time, they emphasize that realizing this potential will require a careful balance between innovation, patient safety, and responsible implementation.
The authors also describe a spectrum of clinical autonomy, running from "advisory" tools that supplement clinical judgment without being embedded in workflows, through "co-pilot" tools that share tasks with a clinician who remains in control, to "navigator" systems that work with minimal human oversight. Almost all AI in clinical use today sits at the advisory and co-pilot end, but as models become able to use multiple data streams simultaneously, “navigator” roles could emerge across personalized treatment, remote patient monitoring, clinical decision support, robotic surgery, and hospital operations.
In doing so, AI may help drive a broader shift toward earlier prediction and prevention of disease, with significant implications for global population health and future therapeutic development.
Senior author Dr Hutan Ashrafian from Imperial College London—who brings expertise across clinical medicine, surgery, business, and innovation—said: “Healthcare systems around the world are facing growing pressures from aging populations, rising rates of chronic disease, and workforce shortages. The need to get this right has never been greater.”
Safe and trusted adoption
Implementing medical AI remains uneven across healthcare systems: while AI has advanced rapidly, most regulatory-approved AI tools are focused on diagnosis and decision support. Most have also been judged on technical accuracy rather than on whether they improve outcomes for patients, and few have been evaluated prospectively once deployed.
Challenges related to clinical validation, integration into existing workflows, data quality, interoperability, energy consumption, and regulatory approval continue to limit widespread deployment. The authors cite AI-based cervical cancer screening as an example: although its use is gaining traction, deployment remains inconsistent, held back by limited clinical validation, variation in clinical practice, and regulatory considerations.
They say that realizing AI's full potential will require advances not only in algorithms and computing power, but also in data infrastructure, workforce readiness, clinical integration, and governance.
Lead author Dr Ahmad Guni from Imperial College London said: “The potential applications of AI in medicine extend far beyond what we are currently seeing in diagnosis. AI could help us detect disease earlier, tailor treatments more precisely, and shift the focus from treating illness to preventing it in the first place. But technology alone isn’t enough, and translating the promise into real-world impact has proven more challenging than many anticipated.”
Regulation alone, they argue, will not be sufficient to ensure impactful adoption. Addressing practical considerations—including reimbursement models, procurement pathways, institutional governance, and ongoing post-deployment monitoring—will be critical. As AI becomes more deeply embedded in clinical care, clear lines of responsibility, including how liability is shared between clinicians, AI developers, and hospitals, and a strong understanding of how clinicians and AI systems can work together effectively will be essential.
Dr Ashrafian said: “AI is attracting so much attention because it offers the possibility not only of improving clinical care, but also of reducing errors and workflow inefficiencies, easing pressures on healthcare services, and ultimately delivering much better experiences and outcomes for patients. Today’s challenge is not simply developing more and more powerful AI, but ensuring that these technologies are integrated in ways that genuinely improve the quality of care for patients.”
Ultimately, the authors conclude that bridging the gap between technical capability and real-world clinical benefit demands sustained investment not only in AI innovation, but also in the governance, infrastructure, sustainability, and trust needed to support its safe and equitable adoption.
Dr Guni said: “To deliver real benefits for patients, we need strong evidence, careful oversight, and close collaboration between clinicians, researchers, regulators, and technology developers. If we get that balance right, AI could help create a more proactive, efficient, and equitable healthcare system for everyone.”
Frontiers in Science
Literature review
Not applicable
Large language medicine: defining a new paradigm in human health
1-Sep-2026