Abstract
Static, pre-compiled software applications can be rigid, limiting their ability to provide dynamic, personalized user experiences and often requiring slow, resource-intensive update cycles. A system is described that can enable software to adapt its functionality or user interface at runtime. The system may utilize an architecture of artificial intelligence (AI) agents to generate code or configuration modifications based on high-level goals. A feature of this system can be an automated validation loop where an AI-based judge evaluates these modifications against a governance corpus of rules for security and correctness. Approved modifications can then be executed within a sandboxed environment in the host application. This framework may facilitate automated software adaptation while providing a mechanism for oversight and risk mitigation associated with executing dynamically generated code.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Leung, Mira, "AI-Driven Runtime Software Adaptation with AI Judge Validation and Sandboxing", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/10693