SafetyDrift: Predicting When AI Agents Cross the Line Before They Actually Do
When an LLM agent reads a confidential file, then writes a summary, then emails it externally, no single step is unsafe, but the sequence is a data leak. We call this safety drift: individually safe actions compounding into violations. Prior work has measured this problem; we predict it. SafetyDrift models agent safety trajectories as absorbing Markov chains, computing the probability that a trajectory will reach a violation within a given number of steps via closed form absorption analysis. A consequence of the monotonic state design is that every agent will eventually violate safety if left unsupervised (absorption probability 1.0 from all states), making the practical question not if b...
| Authors | Aditya Dhodapkar, Farhaan Pishori |
|---|---|
| Primary category | cs.CR |
| Published | 28 March 2026 |
| Last updated at source | 28 March 2026 |
Models and datasets referencing this paper
- aditya593/SafetyDrift-tracesDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.