November 2026

Preparing for Disruptions in AI-Assisted Clinical Care

November 2026 | Clearwater Healthcare Cyber Briefing

This session examined how healthcare organizations can prepare for disruptions involving AI-assisted clinical technologies, including systems that remain operational but produce unreliable information that could affect patient care.

Executive Summary

As artificial intelligence becomes embedded in clinical workflows, healthcare organizations face a new challenge: preparing for situations in which AI systems become unavailable, are compromised, or continue operating while producing clinically unreliable results.

In Clearwater’s November Healthcare Cyber Briefing, Dave Bailey, VP of Consulting Solutions & Strategy, opens with the latest healthcare cybersecurity threats, ransomware activity, breach statistics, and regulatory developments. He then introduces the Health Sector Coordinating Council’s (HSCC) newly published Playbook for AI Clinical Resilience, which he co-authored.

Dave is joined by Dr. Benoit Desjardins, Professor of Radiology at the University of Montreal, Chief Medical Cybersecurity Officer (CMCO) at CHUM, and Chief Medical Information Officer (CMIO) in Radiology at CHUM.

Drawing on clinical, cybersecurity, and AI expertise, they examine how healthcare organizations can identify AI dependencies, recognize unreliable outputs, maintain patient care during disruptions, and establish safe recovery procedures.

The session’s central message: AI clinical resilience requires more than restoring technology. Organizations must recognize when AI can no longer be trusted and maintain safe patient care until its reliability has been demonstrated.

In This Discussion

This briefing explores questions including:

What Changed in Healthcare Cybersecurity Threats in October?

Dave Bailey opens with Clearwater’s Critical Threat Brief, highlighting AI-enabled exploitation, ransomware activity, healthcare breach reporting, and HIPAA enforcement developments.

AI-assisted vulnerability discovery and automated exploitation are increasing pressure on security teams to identify and remediate vulnerabilities quickly. Dave also discusses emerging concerns about AI capabilities being used against operational technology and industrial control environments.

Ransomware activity remains elevated. According to Clearwater’s threat monitoring presented during the briefing, October was the busiest month of 2026 for ransomware leak-site postings involving U.S. healthcare and pharmaceutical organizations, with 78 claimed victims.

Healthcare breach reporting also remained significant. Dave cites 615 reported breaches affecting approximately 87.7 million individuals through October 8, based on HHS OCR breach portal data.

On enforcement, Dave highlights OCR’s October action involving Lehigh Valley Health Network, reinforcing the importance of thorough HIPAA Security Rule risk analysis, information system activity reviews, and appropriate breach notification.

These developments reinforce the need to connect cybersecurity risk management with clinical operations and patient safety.

What Is the HSCC Playbook for AI Clinical Resilience?

Published in October 2026, the Health Sector Coordinating Council’s Playbook for AI Clinical Resilience provides guidance for developing response procedures that protect patients when AI-enabled clinical technologies fail or become compromised.

Dave Bailey, a co-author of the playbook, explains that AI disruptions can result from malicious activity or ordinary technology failures.

Model drift, integration problems, and unreliable outputs can create clinical consequences even when no cybersecurity incident has occurred.

The playbook connects AI risk assessment with clinical safety, incident response, and continuity of care.

Its guidance emphasizes five priorities:

The discussion emphasizes that healthcare organizations should tailor these procedures to their actual clinical workflows rather than rely exclusively on enterprise-wide downtime plans.

How Can AI Become Clinically Unreliable Without Going Offline?

Dr. Benoit Desjardins explains that one of the most concerning AI failure scenarios is not a complete system outage.

Traditional technology failures are often visible. An application becomes unavailable, a computer stops functioning, or a network connection fails.

AI presents a different challenge.

Dr. Benoit Desjardins: “With AI, it’s different, because AI can remain operational”

An AI system can continue producing outputs while becoming clinically unreliable.

Potential causes include model drift, data quality problems, integration failures, software updates, unauthorized modifications, and adversarial attacks.

This creates particular risk when clinicians have developed confidence in a tool after months of reliable performance.

The discussion emphasizes that healthcare organizations must monitor clinical reliability, not simply system availability.

Why Is Mapping AI Across Clinical Workflows Essential?

Healthcare organizations cannot prepare for AI disruptions without understanding where AI is used and how clinical teams depend on it.

Dr. Desjardins describes radiology as an important example because AI is increasingly integrated into imaging interpretation, prioritization, workflow management, and other clinical processes.

A disruption affecting one AI component may influence multiple stages of care.

Organizations should document not only individual AI applications but also their connections to clinical systems, data sources, and workflows.

An effective AI inventory should identify which clinical processes depend on AI, what could happen if those capabilities fail, and which alternative workflows must remain available.

How Can Healthcare Organizations Detect Unreliable AI Outputs?

An AI system that continues generating plausible but incorrect information can be more difficult to identify than one that stops functioning entirely.

Dr. Desjardins emphasizes the importance of continuous clinical performance monitoring and human oversight.

Dr. Benoit Desjardins: “You need a human in the loop.”

Human review is especially important during initial implementation, when organizations are establishing confidence in a system’s performance.

However, oversight should not disappear once an AI application demonstrates consistent reliability.

Healthcare organizations need mechanisms for clinicians to identify unexpected AI behavior, report concerns, and escalate potential patient safety issues.

The discussion reinforces that clinical judgment remains essential even when AI systems have performed reliably over time.

What Happens When AI-Assisted Clinical Prioritization Becomes Unreliable?

Dave Bailey introduces a hypothetical emergency department scenario involving AI-assisted imaging prioritization.

In the scenario, a radiologist notices that potentially urgent cases appear to be ranked incorrectly.

The example illustrates how AI-related disruptions may first be identified by clinicians rather than cybersecurity monitoring tools.

Healthcare organizations need procedures for escalating concerns, evaluating patient safety implications, and determining whether the affected AI system should remain operational.

Response decisions require collaboration among clinical leaders, cybersecurity professionals, IT teams, and organizational leadership.

The panel emphasizes that organizations must consider both the risks of continuing to use unreliable AI and the consequences of disabling technology that has become integrated into clinical workflows.

How Can Healthcare Organizations Maintain Patient Care During AI Disruptions?

A central theme of the webinar is the importance of developing and practicing alternative clinical workflows.

Dr. Desjardins explains that radiology departments already prepare for technology disruptions, including situations where imaging systems or network connections become unavailable.

Depending on the circumstances, teams may need to return to manual processes.

AI introduces additional complexity because clinical workflows may become increasingly dependent on its functionality.

Dr. Benoit Desjardins: “These alternate workflows have to be in the background, have to be available, have to be practiced”

Organizations should determine whether AI components can be disabled independently, whether manual processes remain practical, and how patient care would continue during a prolonged disruption.

Fallback procedures must be tested before they are needed.

When Is It Safe to Restore AI to Clinical Use?

Restoring an AI system requires more than confirming that it is technically operational.

Healthcare organizations must determine whether the system is clinically reliable.

The discussion emphasizes that AI failures may result from cybersecurity incidents, model drift, software defects, or integration problems. Each may require a different investigative and recovery approach.

Dr. Desjardins cautions that immediately disabling AI may not always be the safest response when the technology has become deeply integrated into clinical workflows.

Maintaining modular systems can help organizations isolate problematic AI components without unnecessarily disrupting broader operations.

Vendor communication is also essential. Healthcare organizations should understand when AI models or software are updated and independently evaluate performance following significant changes.

Key Takeaways

Questions This Discussion Answers

About This Healthcare Cyber Briefing

Clearwater’s Healthcare Cyber Briefing is a monthly webinar series examining developments in healthcare cybersecurity, privacy, compliance, and enterprise risk management.

In this November session, Dave Bailey delivers Clearwater’s Critical Threat Brief before introducing the HSCC Playbook for AI Clinical Resilience. He is joined by Dr. Benoit Desjardins of CHUM and the University of Montreal to discuss clinical AI dependencies, unreliable outputs, alternative workflows, and safe recovery procedures.

Together, they provide healthcare executives, CISOs, clinical leaders, IT professionals, and risk management teams with practical guidance for preparing for AI disruptions and maintaining safe patient care.