# Deceptive Alignment Monitoring

Preprint · Andres Carranza, Dhruv Pai, Rylan Schaeffer, Arnuv Tandon, Sanmi Koyejo

- Archive id: arxiv:2307.10569
- Published: 2023-07-20
- arXiv: https://arxiv.org/abs/2307.10569
- Citations: 4
- Source URL: https://arxiv.org/abs/2307.10569
- PDF: https://arxiv.org/pdf/2307.10569
- Provenance: arxiv

## Abstract

As the capabilities of large machine learning models continue to grow, and as the autonomy afforded to such models continues to expand, the spectre of a new adversary looms: the models themselves. The threat that a model might behave in a seemingly reasonable manner, while secretly and subtly modifying its behavior for ulterior reasons is often referred to as deceptive alignment in the AI Safety & Alignment communities. Consequently, we call this new direction Deceptive Alignment Monitoring. In this work, we identify emerging directions in diverse machine learning subfields that we believe will become increasingly important and intertwined in the near future for deceptive alignment monitoring, and we argue that advances in these fields present both long-term challenges and new research opportunities. We conclude by advocating for greater involvement by the adversarial machine learning community in these emerging directions.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [AI Alignment & Safety](https://space-kitty.com/topics/alignment-safety)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.