Advancing Safe, Secure, and Trustworthy AI in Digital Health

Artificial intelligence is rapidly transforming healthcare delivery, clinical decision-making, biomedical research, public health, and patient engagement. The growing use of generative AI, foundation models, and autonomous or semi-autonomous AI agents creates opportunities to improve healthcare accessibility, efficiency, personalization, and quality. However, these systems also introduce risks related to privacy leakage, insecure model training and inference, adversarial manipulation, unreliable outputs, unsafe agent behavior, bias, misinformation, and inadequate oversight. Addressing these challenges requires integrated technical, clinical, ethical, regulatory, and human-centered approaches that protect patients and health data while supporting responsible AI innovation.The goal of this Research Topic is to advance the development, evaluation, deployment, and governance of safe, secure, privacy-preserving, and trustworthy AI systems for digital health.This collection invites submissions that:- advance the safety, security, privacy, reliability, and trustworthiness of AI-enabled digital-health systems;- develop and evaluate secure AI models, privacy-preserving learning, and secure inference methods for healthcare applications;- address risks associated with generative AI, foundation models, and autonomous or semi-autonomous healthcare AI agents;- investigate adversarial robustness, model misuse, data leakage, unreliable outputs, and post-deployment failures;- Examine how AI safety and security affect patient safety, clinical reliability, equity, professional practice, and public trust;- propose governance, assurance, standards, and human-oversight frameworks for responsible healthcare AI;- develop approaches for detecting and mitigating AI-generated health misinformation, disinformation, and misleading content.Suitable themes for manuscripts include (but are not limited to):• Secure AI model development, deployment, and inference in digital health.• Privacy-preserving AI, including federated learning, differential privacy, homomorphic encryption, and confidential computing.• AI safety, robustness, validation, uncertainty assessment, and post-deployment monitoring.• Security and governance of generative AI, foundation models, and healthcare AI agents.• Adversarial threats, including prompt injection, data poisoning, model extraction, and misuse.• Human oversight, explainability, accountability, fairness, and trust calibration.• Health misinformation, disinformation, provenance, and content authenticity.• AI assurance, standards, certification, regulatory compliance, and future-ready protections such as post-quantum security. Topic editors: Dr. Jasleen Kaur, Di Jiang, KORAY KARABINA Source: https://www.frontiersin.org/research-topics/84027/advancing-safe-secure-and-trustworthy-ai-in-digital-health