Healthcare deserves more from AI. With one-quarter of healthcare workers considering leaving their profession each week and a projected global shortage of 10 million workers by 2030, the industry desperately needs effective support. Yet, the low barrier to AI development has led to a flood of poorly conceived tools that add to providers’ workload rather than easing it. Instead of alleviating strain, many AI solutions prove expensive, error-prone, and ill-suited to the complexities of healthcare – creating more problems than they solve.
Recent headlines have reignited debate over AI’s societal value and the risks of accelerated adoption. While the launch of accessible new models like DeepSeek have brought benefits to many, healthcare demands a different calculus – one where safety must take precedence over pure scale. The reality of general AI adoption in this high-stakes sector is sobering: YouGov data shows that one-third of clinicians using AI at work now spend up to three hours a week correcting its mistakes, turning a supposed aid into yet another burden. Skepticism runs deep, with the same YouGov report showing that 52 percent of doctors lack confidence in current AI solutions. The result? A growing state of “pilot paralysis,” where promising trials stall under the weight of accuracy, cost, and integration challenges.
Consider this: Would you use a Swiss army knife for surgery, or trust a medical student for highly complex procedures? Healthcare requires specialized tools for specialized needs. While ChatGPT has accelerated global AI adoption, healthcare cannot rely on AI systems trained on everything from society columns to social media. If these tools fail to deliver, doctors will abandon them – leaving pilots halted in the adoption phase. This cycle of implementation and abandonment doesn’t just waste resources; it erodes trust in the very technology that could revolutionize healthcare delivery.
A recent study by Stanford School of Medicine found that 8-10 percent of patient replies generated by GPT-4 were potentially risky to patients as they weren’t rooted in case facts. Healthcare is heavily regulated for a reason – people’s lives are at risk. Patients deserve knowledgeable doctors, and doctors need AI they can trust to enhance care while ensuring compliance. General-purpose AI, stretched too thin, struggles to grasp the nuanced demands of medicine – from compliance to clarity and context – and safety above all else.
Access to specialized medical foundation models is essential to overcoming pilot paralysis and ensuring consistent, high-quality patient care. While three-quarters of healthcare professionals support AI in practice, their concerns are valid – more than half cite AI errors as their biggest worry. To truly benefit healthcare, AI must pivot to foundation models designed specifically for medicine, trained on real-world patient interactions and peer-reviewed research rather than general internet data. These specialized models will far surpass general-purpose solutions. It’s time to move beyond endless pilots and deploy AI that truly meets the demands of modern medicine. The future of healthcare depends not just on having AI, but on having the right AI – one that doctors and patients can trust, one that enhances efficiency without compromising safety, and one that ultimately drives better outcomes for all.
Andreas Cleve
Andreas Cleve is Corti’s co-founder and CEO. After spending nearly a decade working as a multi-entrepreneur in AI, Andreas founded Corti with Lars Maaløe, pioneering a safe and effective Generative AI platform for healthcare. Corti is a research and development company that specializes in state-of-the-art AI foundation models for healthcare. Corti's mission is to eliminate administrative hurdles in healthcare and life sciences and bring expert-level healthcare reasoning to every corner of the globe, driving down costs and improving the quality of care.
Frederik Brabant, MD
Frederik Brabant, MD is Corti’s Chief Medical Strategy Officer where he shapes the company's medical strategy and expands its global footprint. Previously, he served as the chief medical information officer and VP of international healthcare partnerships at Microsoft Nuance. With nearly a decade of leadership experience at Nuance, Frederik played a pivotal role in driving the success of flagship products such as Dragon Dictation and leading global partnership initiatives worth over millions.
