Hyperchaotic Agentic AI Framework for Image Encryption

This paper presents Agentic Hyperchaos which is an intelligent image encryption model that combines autonomous AI-controlled decision regulation with a 5-dimensional Ikeda hyperchaotic map. Otherwise, in contrast to the traditional chaos-based schemes where the parameterization of the scheme is determined by the fixed parameter settings, the framework proposed introduces an agentic orchestration layer where the parameters of the Ikeda-map and the encryption intensity are dynamically controlled based on the chosen security policy and runtime considerations. High entropy, close to zero-pixel correlation, high NPCR and UACI, as well as good key sensitivity are experimentally verified. The resulting modular architecture offers a scaled foundation of next-generation adaptive visual cryptosystems.