# AI Safety: Not Optional, Not Later

arXiv: https://arxiv.org/abs/2609.10630 · HTML: https://mlsanity.com/paper/2609.10630/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Sep 2026.

## Built on by (most cited first, 1 of 1)

- [A Reference Architecture for Integrating Agentic AI into Enterprise Software: Responsibilities, Interactions, and Variability](https://mlsanity.com/paper/2610.14818/) (2026)

## Abstract (arXiv)

Incidents show that AI safety failures often arise across multiple layers. We present a safety-by-design assurance architecture combining model-level supervision, such as Scientist AI, with system-level controls over scaffolds and harnesses, independent verification, monitoring, and evidence infrastructure, supported by governance for accountability and evidence interoperability.