Something has shifted in the relationship between doctors and artificial intelligence. For years, AI in healthcare was a promise more than a practice, a technology being piloted in research hospitals and discussed at conferences while most clinicians kept working with the same systems they had always used. In 2026, that story has changed. The tools are no longer experimental. They are inside operating rooms, radiology suites, and the administrative back offices of hospitals around the world, and the people using them are reporting results that are harder to dismiss each month.
A global survey published in June 2026 by Philips, drawing on responses from healthcare professionals across 14 countries, found that nearly two-thirds of clinicians have increased their use of AI tools at work. Almost half reported saving at least 132 hours per year as a direct result. Just as telling is what those clinicians say the technology is catching. Thirty-nine percent said AI had identified or prevented a potential medical error at least three times in the past three months alone.
Agentic AI Moves Into the Hospital
The newest wave of healthcare AI goes beyond tools that passively surface information. Agentic AI, systems capable of taking sequences of actions autonomously rather than simply responding to a single query, is being adopted by some of the most prominent health systems in the United States. Mount Sinai Health System in New York and the Mayo Clinic in Minnesota are both using agentic AI to automate repetitive administrative tasks and streamline clinical workflows, freeing staff to focus on direct patient care. According to the same Philips survey, 47 percent of respondents said they are already using or actively assessing AI agents in their organizations.
The ambition behind this shift is substantial. If agentic systems can handle appointment scheduling, prior authorization paperwork, and routine documentation at scale, the time that clinicians recover could be redirected toward the kinds of patient interactions that require human judgment and empathy. The technology is not replacing doctors. It is, at least in principle, returning them to the work they trained for.
Imaging and Drug Discovery Lead the Way
Two areas are seeing particularly concentrated AI investment. In medical imaging, 61 percent of medical technology companies surveyed by Nvidia in its 2026 AI in Healthcare report said they are using AI for radiology and diagnostic imaging. The case for AI here has been building for several years, with studies consistently showing that AI models can match or exceed human radiologists in identifying specific abnormalities in chest X-rays and CT scans, particularly when they are used to flag cases for human review rather than to replace that review entirely.
In pharmaceutical and biotechnology, 57 percent of companies in the same survey said drug discovery is being driven by AI. The timeline from target identification to clinical candidate, a process that once took years of iterative laboratory work, is being compressed as AI models learn to predict how molecules will behave in biological systems. It remains early days, but the first drugs developed with substantial AI involvement are beginning to move through clinical trials.
Data Partnerships and Privacy Questions
Not all of June’s healthcare AI news has been straightforward. The 23andMe Research Institute announced a partnership with HealthEx that would allow users to connect their medical records with their genetic data. The potential scientific value is real. The privacy implications are equally real, raising questions about data governance, patient consent, and national security that regulators are still working through. As AI systems become more capable, the sensitivity of the data they require to function well is increasing in parallel.
The Training Gap That Still Needs Closing
Progress across the sector is genuine, but it is uneven. The same surveys that document AI’s growing impact also document a significant gap in how prepared clinicians feel to use these tools responsibly. Seventy percent of healthcare workers in the Philips survey reported that AI training available to them was inadequate, inconsistent, or simply not available. The skills they most want to develop include checking the accuracy of AI recommendations, understanding legal liability when AI is involved in a clinical decision, and navigating the technical interfaces these systems require.
That gap is not simply an inconvenience. An AI tool that a clinician does not trust or does not know how to interrogate properly is a tool that will either be ignored or used uncritically, neither of which produces the outcomes the technology is capable of. Closing the training gap may prove to be as important as any further advancement in the models themselves.
The direction of travel in healthcare AI is clear. The tools are working, the adoption is accelerating, and the evidence base is growing. The next challenge is making sure that the humans using those tools are as well-prepared as the technology itself.
Stay informed on how AI is transforming healthcare and every other sector at Mylistingo.
