AI Safety: news, advisories and analysis

CyberSecureToday covers AI Safety across 2 articles published between 17 Sep 2026 and 13 Oct 2026. The coverage includes 1 vulnerability advisory and 1 analysis or guidance piece. 1 advisory is rated critical or high severity.

Stories on AI Safety most often intersect with Amazon Web Services, AWS Bedrock and Bedrock Guardrails, with Technology, Cloud & SaaS the most affected sector. The most recent story, “AWS Bedrock Guardrails Vulnerability Enables Adversarial Safety Filter Evasion via Unicode Homoglyph Normalization (CVE-2026-80210)”, was published 13 Oct 2026. Every article is listed below, newest first, with the affected products, severity and remediation guidance in each piece.