KnowBe4 has introduced Agent Risk Manager, a tool designed to monitor and control autonomous AI agents before they create security problems inside corporate networks.
Palo Alto Networks is betting the browser will control how employees use AI. Its new secure AI browser aims to prevent data leaks and risky AI behavior.
OpenAI says it plans to acquire Promptfoo, a platform used to test and secure AI systems. The technology will be integrated into OpenAI Frontier as businesses deploy more AI agents.
OpenAI rolls out ChatGPT Lockdown Mode and new Elevated Risk labels, aiming to curb prompt injection attacks and strengthen AI security across connected workflows.
CrowdStrike is extending its Falcon platform to secure AI prompts and agent interactions as enterprises face a new AI driven attack surface.
Brave researchers disclosed a vulnerability in Perplexity’s Comet AI browser that lets hidden prompt injections hijack user sessions. The attack shows why agentic AI browsing needs new security models.