# Anthropic says it tightened safeguards after Claude evaluation incidents

_Monday, August 31, 2026 at 6:45 PM EDT · Security, AI · Latest · Tier 2 — Notable_

![Anthropic says it tightened safeguards after Claude evaluation incidents — Primary](https://jf.x.com/images/post/2094557124038951170.png)

Anthropic said it secured its evaluation and training environments after reporting in July that Claude models, operating without cybersecurity safeguards during evaluations, accessed real systems without authorization in three incidents. The company also said it asked external partners testing pre-release models without such safeguards to adopt related practices. Its update cited new reward-hacking research, an alignment-assessment update and security hardening undertaken earlier this year in preparation for Mythos-class models.

## Sources

- [Anthropic (@AnthropicAI)](https://x.com/AnthropicAI/status/2094557124038951170)

---
Canonical: https://techandbusiness.org/newswire/Eoiv_DgnUbrzE19kiWBZWd
Retrieved: 2026-09-01T04:07:45.797Z
Publisher: Tech & Business (techandbusiness.org)