By Dr. Ken Washington
During a KPMG-sponsored discussion, Susan Sly and Naresh Reddy explored the realities of deploying AI in healthcare at enterprise scale. Their conversation echoed a recurring theme in business: organizations are moving beyond AI experimentation and are now confronting the practical challenges of governance, trust, and implementation at scale. Success will require more than powerful models; it will require confidence that AI systems are reliable, transparent, scalable, and worthy of adoption.
Susan emphasized that trust is central to bringing an AI healthcare product to market, noting that her company may eventually need medical malpractice insurance. That observation illustrates how AI companies are becoming part of the healthcare delivery ecosystem, where accountability is no longer theoretical but operational.
Once an AI system begins influencing diagnosis, treatment, or patient care, developers must be able to explain how their systems were trained, how conclusions are generated, and what safeguards exist when the technology encounters circumstances outside its intended use.
Ken Washington, Phd
Ai Healthcare
Day one of this conference highlighted a second major challenge: healthcare data remains largely unstructured and isolated in silos. Clinical notes, diagnostic images, laboratory results, pharmacy records, and patient-generated information often reside in disconnected systems.
This is consistent with my own experience.
Artificial intelligence depends on usable, representative, and well-governed data. Yet healthcare information is frequently incomplete, duplicated, or fragmented across providers. An AI system cannot create a complete understanding from an incomplete picture.
Before healthcare can capture the full value of AI, the industry must improve interoperability, data governance, privacy, and responsible information sharing. The challenge is not collecting more data but transforming existing data into trusted and accessible forms.
The strongest message from this kickoff event of Ai4 was that successful healthcare AI companies will not merely build better algorithms. They will build systems that clinicians, patients, regulators, and healthcare organizations trust. In healthcare, trust is not simply another feature—it is the foundation upon which every successful AI solution will ultimately depend.