Artificial intelligence was the central theme at the MedTech 2026 conference organized by AdvaMed, where the discussion moved beyond merely showcasing technical capabilities to delving into deep operational and regulatory challenges. While optimism prevails, industry leaders provided critical insights suggesting that the next phase will transition from "media hype" to confronting the complexities of actual implementation in a highly regulated medical environment.
Challenges of System Management and Regulatory Compliance
Erez Kaminski, CEO of Ketryx, warned that the reality of deploying artificial intelligence is more complex than many imagine. He noted that the level of actual deployment remains below expectations, predicting that the biggest problem will lie in "managing complexity."
According to Kaminski, companies that currently struggle to manage a single regulated IT system will face a massive challenge when they have to manage hundreds of AI agents that are constantly changing, especially since many of these use cases will overlap with sensitive quality and R&D processes.
Specialization Strategy and Data as a Competitive Barrier
For his part, Mick Farrell, CEO of ResMed, focused on the importance of "vertical ownership." Farrell believes that success does not come from trying to apply AI everywhere, but from building infrastructure specialized in specific areas such as respiratory and sleep health.
He explained that the real power lies in leveraging big data that no one else possesses; ResMed holds data covering 28 billion nights, giving it the ability to apply AI at scale to serve millions of patients, which is difficult for competitors to replicate. He also pointed to the use of AI in supply chains to achieve cost savings to combat inflation.
The Gap Between Responsibility and Authority in Intelligent Systems
In the context of cybersecurity and governance, Zahra Timsah, CEO of i-GENTIC AI, raised a fundamental point regarding granting AI responsibilities without clearly defining its authorities. She warned that the system might provide a technically accurate answer, but could commit a grave error by sharing patient information inappropriately, bypassing a regulatory requirement, or acting based on outdated legislation.
Timsah emphasized that this challenge represents a critical intersection between AI policy and cybersecurity, where it should not be limited to defining what should be done (policy), but must control access and verify every action before it is executed (security).
Distinguishing Between Productivity and Superficial Innovation
Mike White, a partner at Alexander Group, called for distinguishing between productive and non-productive uses of AI. He believes that real use cases are those that enhance productivity in business processes, such as engagement agents for qualifying potential customers in sales and marketing teams.
Conversely, White criticized some trends he sees as unfeasible, such as using AI to analyze customers' facial expressions during sales calls to suggest reactions, considering that this type of application lacks real practical value compared to agents that support core operations.
What Does This Mean for the Technical Future in the Medical Sector?
These insights show that the medical technology sector is undergoing a phase of technical maturity; the question is no longer "Can AI do this?" but "How can we manage it and ensure its compliance?" The radical change will occur in how AI is integrated into the product development lifecycle (R&D) and quality processes, which requires technical infrastructure capable of handling hundreds of changing systems, rather than just adding smart features to final products.