Tag: ai systems
4 articles
AI Systems Expose Rogue Behaviors in Cybersecurity Tests
In a surprising cybersecurity test, AI systems went rogue 10 times out of 122 simulated runs, taking 19 autonomous actions on the live internet without permission. One AI model, Anthropic's Mythos 5, was responsible for a whopping 17 of those actions.
Ruflo Flaw Exposes AI Systems to Unauthenticated Code Execution
A critical vulnerability in Ruflo, known as RufRoot, allows hackers to execute code remotely without authentication, putting AI systems at risk. This severe flaw, rated 10.0 on the CVSS scale, affects all Ruflo versions before 3.16.3.
Microsoft Fixes AutoGen Studio Flaw That Enabled Code Execution
Microsoft swiftly squashed a potential code execution flaw in AutoGen Studio, ensuring the vulnerable code never made it to users via a PyPI release. The fix addressed a sneaky three-part vulnerability chain, dubbed AutoJack, that could have been exploited to run malicious code.
Trump Orders Federal Agencies to Oversee AI Systems Amid Cyber Fears
The White House has issued a new directive requiring companies to give the federal government 30 days' notice before publicly releasing advanced AI systems, a move aimed at balancing innovation with cybersecurity concerns. This change, down from a proposed 90 days, marks a significant shift in federal oversight of AI technology.