4+ day, 4+ hour ago   (23+ words) Nearly half of organizations have zero visibility into AI agent traffic. Check Point and Nightfall AI take different approaches to closing the gap....

1+ week, 3+ day ago   (306+ words) OpenAI has committed $1bn in subsidised access to its Daybreak cyber tools for critical infrastructure operators with no security budget, announced the same day it shipped a model meeting its own Critical cyber threshold....

1+ week, 4+ day ago   (818+ words) Booz Allen scored nine American and nine Chinese models on how far each could get through a real intrusion unaided. Only Anthropic's Claude Mythos reached the end. Then the firm gave a model ranked 15th an attack harness, and it matched…...

1+ week, 5+ day ago   (900+ words) Johann Rehberger got Claude Code to run an attacker's code in up to 80% of attempts, starting with a request to summarise a web page. The model's refusal to run a supplied binary is what opened the door. Anthropic told him…...

2+ week, 4+ day ago   (482+ words) The technical report on the Hugging Face breach also finds that OpenAI's own training rewarded agents for exploiting their environment, and Europe's reporting rules may not reach the model that did most of the work OpenAI’s technical report on the…...

2+ week, 5+ day ago   (439+ words) Security firms argue you cannot measure what a model can do without realistic conditions, while Europe already requires serious incidents to be reported without undue delay AI Learning and Artificial Intelligence After AI models from at least three companies reached…...