(Toronto, April 1, 2026) — JMIR Publications has released a significant article addressing the emerging risks associated with autonomous AI systems in clinical environments. Titled “Emerging Risks of AI-to-AI Interactions in Health Care: Lessons From Moltbook,” the piece highlights the challenges posed by high-risk AI agents that begin to communicate directly with one another for triage and scheduling, forming a “digital ecosystem” that can function without active human oversight. The report is authored by Tejas S. Athni, a correspondent for JMIR.
The analysis draws on the 2026 “Moltbook” experiment—a social network designed specifically for AI-to-AI interaction—to illustrate the potential impacts on healthcare. Athni warns that while interconnected systems may enhance operational efficiency, they also introduce a “lethal trifecta” of risks that includes the rapid spread of errors, heightened data leaks, and the unintended emergence of hierarchies among AI agents.
One of the most pressing issues identified in the report is the propagation of errors. In a networked environment, a single misinterpretation by a diagnostic AI can lead to catastrophic outcomes. For instance, if a diagnostic AI mislabels a fracture, this error can be accepted and amplified by downstream AI agents responsible for bed allocation and triage, resulting in systemic medical errors that could compromise patient care.
Furthermore, the interconnected nature of these systems can accelerate the risk of data leaks. Autonomous agents often share and withhold information in ways that their creators may not have anticipated. This vulnerability opens the door for adversarial actors to exploit these pathways, facilitating model inversion or membership inference attacks that could compromise protected health information (PHI) at unprecedented speeds. As such, the implications for patient privacy are significant, raising alarms about the security of sensitive healthcare data.
Another critical issue highlighted in the report is the spontaneous development of hierarchies among AI agents. Observations from the Moltbook experiment suggest that these agents can inadvertently establish dominant or subordinate roles, complicating clinical decision-making. In a hospital context, for example, an AI responsible for ICU allocation could begin to override diagnostic agents, thereby creating de facto priorities that might not align with ethical standards or established clinical protocols.
In light of these challenges, the article advocates for a proactive approach to the design of medical AI systems, urging a shift away from reactive fixes toward “preventive design.” Experts emphasize the need for transparency and robust safeguards as autonomous systems increasingly integrate into healthcare settings. Recommendations include the introduction of human-centric guardrails that require human validation—such as radiologists reviewing AI classifications—before any autonomous decision is enacted.
The report also calls for aggressive stress-testing of AI-to-AI communication protocols, utilizing techniques like red teaming to uncover vulnerabilities prior to deploying these systems in live clinical settings. Additionally, maintaining clear and trackable records of interactions and decisions made by autonomous agents is vital for ensuring accountability within these complex systems.
“The risks of AI-to-AI interactions must be taken seriously as autonomous systems become integrated into healthcare,” Athni concludes. “The Moltbook experiment offers a critical lens to ensure these digital dangers do not translate into real-world patient harm.” This forward-looking perspective underscores the need for ongoing dialogue and research as the healthcare sector grapples with the implications of integrating AI technologies.
Please cite as:
Athni T
Emerging Risks of AI-to-AI Interactions in Health Care: Lessons From Moltbook
J Med Internet Res 2026;28:e96199
URL: https://www.jmir.org/2026/1/e96199
DOI: 10.2196/96199
About JMIR Publications
JMIR Publications is a prominent open access publisher of digital health research, dedicated to advancing the field through a range of peer-reviewed journals, including the Journal of Medical Internet Research. Its mission includes supporting researchers and maximizing the impact of their work.
Media Contact:
Dennis O’Brien, Vice President, Communications & Partnerships
JMIR Publications
+1 416-583-2040
See also
AI Study Reveals Generated Faces Indistinguishable from Real Photos, Erodes Trust in Visual Media
Gen AI Revolutionizes Market Research, Transforming $140B Industry Dynamics
Researchers Unlock Light-Based AI Operations for Significant Energy Efficiency Gains
Tempus AI Reports $334M Earnings Surge, Unveils Lymphoma Research Partnership
Iaroslav Argunov Reveals Big Data Methodology Boosting Construction Profits by Billions
