arXiv Artificial Intelligence

Agent Safety Should Be a Runtime Contract

1 hour ago 13.08.2026 07:00 👁 2 views
Agent Safety Should Be a Runtime Contract

Quick summary

arXiv:2608.11274v1 Announce Type: cross Abstract: The dominant paradigm treats AI safety as a property to be instilled during model training via RLHF, DPO, or Constitutional AI. We argue this is structurally insufficient for autonomous agents that execute code, mutate files, send messages, and modify databases. Agent safety should be a runtime contract enforced by the harness, and the contract has two complementary faces. The preventive face blocks dangerous actions before they happen via sandboxes, permission gates, output filters, and trajectory monitors. The evidential face requires verifia

Key takeaways

  • arXiv:2608.11274v1 Announce Type: cross Abstract: The dominant paradigm treats AI safety as a property to be instilled during model training via RLHF, DPO, or Constitutional AI.
  • We argue this is structurally insufficient for autonomous agents that execute code, mutate files, send messages, and modify databases.
  • Agent safety should be a runtime contract enforced by the harness, and the contract has two complementary faces.

Why it matters

This development is a reminder to test misuse and data-leak scenarios alongside speed and quality. Trust should come from testable controls and clear failure reporting, not protection claims alone.

Kaynak sitede devamını oku: arXiv Artificial Intelligence ↗