Philips CEO: AI's Real-World Impact & Trust Challenge in Healthcare

Artificial intelligence is quietly revolutionizing the efficiency and potential of U.S. healthcare, even as the landscape of government health policy and spending continues to shift dramatically. Philips, the venerable electronics manufacturer that has transformed into a leading medtech provider, is at the forefront of this AI-driven healthcare revolution, actively streamlining and accelerating the workflow of patient care. According to Jeff DiLullo, CEO of Philips North America, technology holds immense power to impact health outcomes today, from optimizing radiology scans to expediting cancer diagnoses. He emphasizes that leaders across industries must rethink traditional approaches to best meet the demands of this evolving moment.

While there is much discussion about AI’s transformative power, its actual implementation and impact sometimes lag behind the hype in various business sectors. However, DiLullo highlights that within medtech, AI’s influence is already substantial and immediate. He points to Philips’ 2025 Future Health Index, which revealed that AI applications in certain healthcare “compartments” or areas are remarkably mature and have even received FDA clearance, deeming them safe for clinical use. In contrast, other areas remain more experimental. A significant hurdle, however, is the nascent “trust factor” among a broader population, which currently represents the biggest barrier to widespread deployment.

This trust gap is evident in the data: approximately 60 to 65% of clinicians express trust in AI, yet only about a third of patients, particularly older individuals, share that sentiment. DiLullo believes that bridging this divide is a collective responsibility, with Philips playing a crucial role in providing validated, FDA-cleared diagnostic capabilities powered by AI. He notes that younger generations, being “digitally fluid,” inherently trust AI models. For older patients, the key lies with the healthcare practitioners themselves. If doctors and nurses believe in the credibility and efficacy of AI—using it to augment their analyses and diagnostics rather than replace them—patient trust will ultimately increase. When clinicians see the value in AI, recognizing its ability to free up more time for patient interaction and reduce their own stress, adoption is expected to accelerate parabolically within health systems in the coming years.

Beyond administrative functions, AI’s practical application for practitioners is already making a tangible difference, particularly in radiology. Early diagnosis often correlates with better outcomes, yet waiting times for scans can be prolonged. DiLullo cites how Philips’ AI-integrated MRI systems significantly reduce scan times, sometimes by half or even two-thirds. A scan that once took 45 minutes can now be completed in just 20 minutes, thanks to “smart speed” technology that compresses scanning time by removing noise rather than filling in blanks. This not only yields a better-quality scan in less time but also allows radiologists to process more studies per day—potentially 20 instead of 12 or 15. This increased throughput translates to more patients served, higher reimbursement for hospitals, and ultimately, improved patient care. Furthermore, AI can pinpoint specific areas in digital images, directing radiologists to critical regions for immediate determination. The entire diagnostic process, including digital pathology for cancer findings, can now be streamlined to mere hours, a monumental shift from previous timelines.

Concerns about “AI hallucinations,” often associated with generative AI models, are less pertinent to the diagnostic AI currently deployed in healthcare, as human oversight remains integral. While caution and robust governance are essential for exploring more advanced generative AI, DiLullo stresses that avoiding experimentation is not an option. He highlights that current AI applications, such as smart speed in radiology workflows, expedited tumor board meetings, and on-demand virtual consultations, are already making a profound impact, though not yet at their full potential. Leading institutions like Massachusetts General Brigham, Stanford, and Mount Sinai are actively leveraging population health data to train AI models for specific and broad use cases, demonstrating the immense, immediate opportunities available.

DiLullo emphasizes that healthcare systems do not need to wait for a “silver bullet” solution that promises eternal life or cures every ailment. Instead, the focus should be on optimizing and enhancing the existing system. Just as one doesn’t immediately drive on the Autobahn when first learning to drive, there is significant work to be done in the “neighborhood” of current healthcare operations. The vast majority—DiLullo estimates 80%—of AI’s potential impact can be realized today by driving productivity at scale with mature AI and virtual capabilities. This immediate, game-changing potential represents the next great opportunity for healthcare delivery, driven by the profound and urgent need for efficiency and improved patient outcomes.