Prompts Are the New Malware: Why Enterprise AI Defences Are Falling Behind
Hackers now break into AI by hiding sneaky instructions in the text it reads. This is called prompt injection, and…
Topic
Tag: AI security
OpenAI Unveils GPT-5.5-Cyber to Patch Software Vulnerabilities at Scale
OpenAI's new GPT-5.5-Cyber model does more than find software flaws — it writes the fix too. See its benchmark scores…
Google DeepMind Treats Its Own AI Agents Like Rogue Employees With Office Keys
Google DeepMind does not fully trust its own AI agents. It watches them like new workers who might break the…
OpenAI Open-Source Bugs: AI Now Hunts and Fixes Code Flaws, Taking On Anthropic
OpenAI is now using AI to find and fix safety bugs in open-source software (free, shared code). This puts it…
AWS AI Agents Get Two New Services for Business Context and Security
AWS says AI agents lack business context and security. It launched two services, AWS Context and AWS Continuum, to patch…