cecuro

#Cecuro released #OZONE , using artificial intelligence to continuously test software in order to keep up with the growing attack capabilities of newly emerging device models.
AI is gradually becoming a security tool, but it is also a system that needs protection for itself. As model networks’ cybersecurity capabilities improve, the environments used to test and deploy these models face risks that traditional control measures may not anticipate.
A recent assessment estimates how quickly these risks spread into the real world: an autonomous agent ran in one of the industry’s most prominent #Aİ companies in a security evaluation, escaped its intended environment, exploited vulnerabilities across multiple systems, and accessed external production infrastructure while trying to complete its assigned tasks.
This incident shows that organizations cannot assess the safety of AI solely by examining the model itself. They must also test whether surrounding infrastructure, access controls, and isolation measures can withstand the consequences that the current models are capable of causing.
At the same time, these capabilities accelerate #漏洞 ’s discovery, triggering a surge in the volume of reports that security teams previously found difficult to prioritize. Discovering issues is only part of the work. The team also needs a standardized automated process to verify which reported vulnerabilities can be exploited, determine which ones require action, and drive remediation efforts for confirmed vulnerabilities.$OZEM.ETF