Check Point researchers tried to break the frameworks enterprises use to build AI apps. Now they're telling Black Hat ...
AI-generated code is accelerating software development, but review processes are not keeping pace. Learn why enterprises need ...
Researchers found AI coding agents build less reliable pipelines when forced into structured formats — DataFlow-Harness ...
Filters don't stop prompt injection; architecture does. A field guide to the lethal trifecta, the rule of two, Dual-LLM and ...
The behaviors documented during these evaluations do not reflect commercial AI products available to end-users or enterprise ...
A zero-day SQL-injection vulnerability in Metabase Cloud is under exploitation in the wild, and it could spell trouble for many downstream organizations. Metabase, which provides AI-driven business ...
Nation-state grade iOS exploit chains Coruna and DarkSword are proliferating among cybercriminals, with 17,000 domains hosting variants.
Nearly 800 malicious npm packages deliver a cross-platform RAT and infostealer, using WEL1DROPPER to target Windows, macOS, ...
MAESTRO analysis of OpenAI and Anthropic incidents reveals how per-layer failures, not just the model, shaped security outcomes.
Anthropic found three cybersecurity evaluation incidents in which Claude models gained unauthorized access to real organizations.
Las Vegas was hotter than hell last week, but not as hot as the market for artificial intelligence-enabled security at Black ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results