AI is now embedded in critical digital infrastructure and that brings serious risk. Machine learning models perform well, but their opacity and vulnerability to adversarial attack make them poorly suited to high-stakes cybersecurity environments where decisions must be explainable and verifiable.

Formal logic and automated reasoning offer a rigorous path forward. Grounded in mathematical foundations for knowledge representation and deductive inference, these methods provide the tools needed to verify and validate AI behaviour in ways that purely statistical approaches cannot. Hybrid architectures that combine symbolic reasoning with statistical learning can enforce security policies, satisfy safety constraints, and produce outputs that humans can audit. As cyber threats grow more sophisticated, this shift toward logic-grounded AI is not optional; it is a design requirement for systems we can genuinely trust.

This Research Topic addresses a core tension in modern AI: the gap between predictive power and verifiable behaviour. We cannot afford to deploy autonomous systems in critical infrastructure if we cannot explain their decisions or prove their safety properties. Our goal is to advance hybrid frameworks that bring together the empirical strengths of statistical learning and the rigour of symbolic inference. By integrating formal logic into AI architectures, we aim to produce systems that operate within interpretable parameters, adhere to security constraints, and can be formally verified. The result should be a principled foundation for trustworthy AI that holds up under adversarial conditions.

This Research Topic invites submissions that bridge formal methods and modern AI, with a focus on cybersecurity applications. We are particularly interested in work that advances the verifiability, transparency, and resilience of intelligent systems.

Key themes include:

• Neuro-symbolic architectures for threat intelligence
• Formal verification and validation of deep learning models
• Logic-based explainability in high-stakes decision-making
• Automated reasoning for security policy enforcement
• Provably secure AI frameworks

We welcome Original Research, Systematic Reviews, Methods papers, and Perspective articles. Both theoretical contributions and practical applications are in scope. We look forward to work that helps define what trustworthy AI looks like at the intersection of logic and security.

We welcome a diverse range of manuscript types, including Original Research, Systematic Reviews, Methods papers, and Perspective articles that offer novel theoretical insights or practical applications. By uniting these disciplines, we aim to establish a principled roadmap for Trustworthy AI that can withstand the adversarial nature of contemporary cyber landscapes. We look forward to contributions in shaping this critical intersection of logic and security.

Are you interested in contributing a manuscript to this Research Topic?

Logic-Centric Reasoning Frameworks for Cybersecurity and Trustworthy AI Systems