By Editorial TeamDespite the undeniable potential of artificial intelligence to revolutionize healthcare, a significant chasm persists between technological advancement and clinical integration. Recent intelligence reveals a stark reality: only 26% of clinicians currently trust AI for critical clinical decision-making, as highlighted by CW3-DP-09. This trust deficit emerges as the primary barrier to widespread AI adoption, even as innovative solutions from companies like Viz.ai, Aidoc, and Epic Systems push the boundaries of what’s possible.
The Trust Gap: A Critical Barrier to AI Adoption
The reluctance of a vast majority of clinicians to fully embrace AI in their daily practice underscores a fundamental challenge in the digital transformation of healthcare. While the promise of AI to enhance diagnostic accuracy, streamline workflows, and personalize treatment plans is clear, the path to achieving these benefits is fraught with hurdles. This trust gap, identified as the number one adoption barrier, is not merely a perception issue but a complex interplay of factors ranging from explainability to regulatory uncertainty.
Thought leaders like Robert Wachter have consistently emphasized the need for AI to augment, rather than replace, human intelligence in healthcare. His recent book, “A Giant Leap: How AI is Transforming Healthcare and What That Means for Our Future,” published in 2026, further explores how AI can improve care, reduce burnout, and help repair a broken healthcare system, while also highlighting the unprecedented speed of AI progress and the importance of clinician engagement. Similarly, Eric Topol has championed the concept of “high-fidelity” AI, arguing that trust will only be built when AI tools demonstrate consistent, verifiable accuracy and clinical utility, freeing up clinicians for more meaningful patient interactions rather than adding to their cognitive load. His latest book, “Deep Medicine,” published in 2026, continues to support the rise of AI in healthcare for enhancing the human side of care. While a March 2026 AMA survey indicated that 72% of doctors are using generative AI for at least one use case, and 35% for direct patient care, Topol notes there is still limited evidence for LLMs benefiting patients or doctors for health outcomes beyond administrative tasks. Mark Sendak, focusing on implementation science, often highlights that successful AI adoption hinges on understanding and addressing the practical concerns of end-users within their specific clinical contexts. Sendak is also the co-founder and CEO of Vega Health, a company launched in October 2025 to help health systems purchase and monitor AI solutions.
Companies at the forefront of healthcare AI are grappling with this trust deficit. Viz.ai, for instance, has demonstrated success in acute care settings, particularly in stroke detection and triage, by providing rapid, actionable insights. In 2026, Viz.ai launched Viz Pulmonary Suite, the first comprehensive AI-powered solution dedicated to pulmonary care delivery, and expanded into neurodegenerative disease through a collaboration with Cortechs.ai. The company also launched Viz Agent Studio to enable health systems to build and scale customizable care pathways and announced a strategic collaboration with Johnson & Johnson to expand access to its Subdural Hemorrhage software solution. Viz.ai’s solutions are now used in over 2,000 hospitals in the US and EMEA. Aidoc offers AI solutions for radiology, aiding in the detection of critical conditions across various imaging modalities. In January 2026, Aidoc received FDA clearance for CARE, a comprehensive foundation model AI with 11 newly cleared indications, bringing its total to 14 indications for triage in emergency departments and ambulatory settings. In April 2026, Aidoc raised $150 million in Series E financing to advance its AI models and expand globally, and in June 2026, it received FDA Breakthrough Device Designation for “First Read,” an AI feature designed to analyze chest X-rays and generate draft report text. Both companies have focused on creating tools that integrate seamlessly into existing clinical workflows, aiming to prove their value through tangible improvements in patient outcomes and operational efficiency. However, even with proven efficacy, the broader adoption is constrained by the overarching trust issue.
The challenge extends to broader platforms as well. Epic Systems, a dominant force in electronic health records (EHRs), is actively integrating various AI capabilities directly into its ecosystem. By March 2026, over 85% of Epic’s customers were actively using its AI tools, including Art for clinical applications, Penny for financial operations, and Emmie for patient interactions. In February 2026, Epic rolled out AI Charting, a built-in ambient AI scribe tool. These EHR-integrated AI solutions aim to embed AI insights directly into the clinician’s workflow, from predictive analytics for patient deterioration to automated documentation. Epic also teased its future AI roadmap at the HIMSS 2026 conference, including an “Agent Factory” for building and monitoring AI agents. While this integration offers unparalleled access, the core issue of clinician trust in these embedded AI recommendations remains paramount. The sheer volume and complexity of data processed by these systems can sometimes create a “black box” perception, hindering user confidence.
Navigating the Regulatory Landscape and Building Confidence
The regulatory environment plays a crucial role in shaping clinician trust and, consequently, AI adoption. The FDA’s Software as a Medical Device (SaMD) Framework provides a pathway for the regulation of AI tools that function as medical devices, ensuring they meet standards for safety and effectiveness. By early 2026, the FDA had authorized over 1,350 AI-enabled devices, doubling the number from 2022. The FDA’s Clinical Decision Support Software Guidance was updated in January 2026, and the new Quality Management System Regulation (QMSR) took effect in February 2026, aligning FDA requirements with ISO 13485:2016 standards. The FDA has also issued final guidance establishing Predetermined Change Control Plans (PCCPs) for managing modifications to AI/ML-enabled devices across the total product lifecycle. Companies like Viz.ai and Aidoc have successfully navigated this framework, obtaining clearances that lend a degree of credibility to their offerings. However, the framework is continually evolving, particularly concerning adaptive AI models, and the transparency around these regulatory processes is key to fostering trust.
Further shaping the landscape is the ONC’s HTI-1 (Health Data, Technology, and Interoperability: Certification Program Updates, Algorithm Transparency, and Information Blocking) rule. This regulation emphasizes transparency for AI/ML-based clinical decision support tools, requiring developers to provide clear, accessible information about the AI’s training data, intended use, and performance. This push for transparency directly addresses some of the “black box” concerns that contribute to clinician skepticism, aligning with the broader goal of building trust. ONC HTI-1 rule details
Major healthcare organizations are also actively engaged in understanding and influencing AI adoption. The American Medical Association (AMA) has consistently advocated for ethical AI development and deployment, emphasizing the need for clinician input and oversight. Their guidelines often focus on ensuring AI tools enhance, rather than diminish, the physician-patient relationship. A March 2026 AMA survey found that 72% of doctors are using generative AI for at least one use case. KLAS Research, through its extensive surveys and reports, offers valuable insights into the real-world performance and clinician satisfaction with various health IT solutions, including AI. Their findings frequently highlight the critical link between usability, effectiveness, and trust. The American Hospital Association (AHA) and HIMSS (Healthcare Information and Management Systems Society) similarly contribute by fostering discussions, developing best practices, and advocating for policies that support responsible AI integration within health systems. In May 2026, the AHA launched a three-year initiative to scale AI and telehealth adoption at hospitals, and in February 2026, they provided recommendations to HHS on reducing regulatory barriers and ensuring safe AI use. HIMSS Global Health Conference & Exhibition in March 2026 and the HIMSS AI in Healthcare Forum in June 2026 both featured extensive discussions on AI. These organizations collectively underscore that technical prowess alone is insufficient; successful adoption requires a holistic approach that prioritizes trust, transparency, and clinical utility. AMA ethical AI principles
The Path Forward: From Skepticism to Seamless Integration
The finding that only 26% of clinicians trust AI for clinical decisions presents a significant challenge but also a clear roadmap for the future of healthcare AI. The key takeaway is that the “trust gap” is not an insurmountable obstacle but a direct call to action for developers, regulators, and health systems. Companies like Viz.ai, Aidoc, and Epic Systems, along with the broader ecosystem of EHR-integrated AI solutions, must continue to prioritize explainability, validate clinical utility with robust evidence, and ensure their products are designed with the clinician’s workflow and patient safety at the forefront. Aidoc’s recent FDA clearances for its foundation model AI and Breakthrough Device Designation for AI-powered draft reporting, alongside Viz.ai’s expansion into new clinical areas and Epic’s widespread AI adoption and new AI Charting tool, demonstrate significant advancements in this direction. Regulatory bodies, through frameworks like FDA SaMD, which has seen over 1,350 authorized AI-enabled devices by early 2026 and updated guidance on CDS and PCCPs, and initiatives like ONC HTI-1, are pushing for greater transparency and accountability, which are foundational to building confidence. As Robert Wachter, Eric Topol, and Mark Sendak have articulated, the future of AI in healthcare hinges on its ability to genuinely empower clinicians, not complicate their practice. Investment trends and regulatory scrutiny will increasingly favor companies that can demonstrate not just technological innovation, but also a deep understanding of human factors and a commitment to earning and maintaining clinician trust. KLAS Research AI adoption report
Frequently Asked Questions
What is the current level of clinician trust in AI for critical clinical decision-making?
Only 26% of clinicians currently trust AI for critical clinical decision-making. This trust deficit is identified as the primary barrier to widespread AI adoption in healthcare.
What are some of the key factors contributing to the ‘trust gap’ in AI adoption among clinicians?
The trust gap is a complex interplay of factors including explainability of AI models and regulatory uncertainty. This reluctance is not merely a perception issue but a fundamental challenge in the digital transformation of healthcare.
How are leading healthcare AI companies addressing the clinician trust deficit?
Companies like Viz.ai and Aidoc are focusing on creating tools that integrate seamlessly into existing clinical workflows and demonstrate tangible improvements in patient outcomes and operational efficiency. Epic Systems is integrating AI capabilities directly into its EHR ecosystem to embed AI insights into the clinician’s workflow.
What is the role of regulatory bodies like the FDA in building confidence in AI tools?
The FDA’s Software as a Medical Device (SaMD) Framework provides a pathway for regulating AI tools that function as medical devices. This framework ensures that AI tools meet standards for safety and effectiveness, which can help build clinician confidence.
