TL;DR
The paper introduces Arbiter-K, a governance-first execution architecture with a semantic ISA that enhances security and robustness in agentic AI systems by actively intercepting unsafe actions and enabling autonomous corrections.
Contribution
It proposes a novel neuro-symbolic kernel with a semantic ISA for agentic AI, improving security and control over traditional orchestration methods.
Findings
Achieves 76% to 95% unsafe interception rates.
Enforces security as a microarchitectural property.
Demonstrates 92.79% absolute gain over native policies.
Abstract
The transition of agentic AI from brittle prototypes to production systems is stalled by a pervasive crisis of craft. We suggest that the prevailing orchestration paradigm-delegating the system control loop to large language models and merely patching with heuristic guardrails-is the root cause of this fragility. Instead, we propose Arbiter-K, a Governance-First execution architecture that reconceptualizes the underlying model as a Probabilistic Processing Unit encapsulated by a deterministic, neuro-symbolic kernel. Arbiter-K implements a Semantic Instruction Set Architecture (ISA) to reify probabilistic messages into discrete instructions. This allows the kernel to maintain a Security Context Registry and construct an Instruction Dependency Graph at runtime, enabling active taint propagation based on the data-flow pedigree of each reasoning node. By leveraging this mechanism, Arbiter-K…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
