Anthropic’s AI model Claude bypassed paywall restrictions and accessed Georgia state’s voter data during an internal statistics project ahead of the midterm elections.
Cyber Security
Cyber criminals, state-sponsored hackers and even the occasional disgruntled employee are constantly looking to gain unauthorized access for a variety of purposes: theft of money, cyber espionage, personal information for sale or for use in scams, and damage to critical infrastructure for just a few of the most common.
So how does an organization mitigate an entire world full of continual cyber attacks? Just as buildings have a number of necessary elements of physical security: access control, cameras, alarms and so on; there are similar key elements of cyber security that are absolutely vital for just about any modern business.
It starts with identifying and closing the most common doors that attackers use. For example, phishing attacks on employees are far and away the most common initial point of entry. The breach of even a low-level employee account can quickly turn into an escalation in access privileges and the ability to reach sensitive information. This is also true of smart devices, which are generally more poorly secured than computers and phones.
A healthcare data breach has hit laboratory giant Quest Diagnostics Incorporated, exposing “personal information and patient records” and prompting an investigation.
Major U.S. banks are blocking third-party AI agents from accessing customer accounts amid concerns that they could take actions customers did not request.
OpenAI says its AI agents disabled safety monitoring guardrails during routine tasks after misinterpreting instructions as a directive to do so, with one agent even lying about it.
CISA and partner agencies warn that SafePay ransomware has compromised over 600 critical infrastructure organizations across the Americas, Europe, and Australia since 2024.
Organizations that build visibility and connectivity discipline into how they deploy agents now won’t be the ones explaining a preventable breach twelve months from now.
New government guidance for AI incidents is directed at organizations that train and evaluate high-capability frontier AI models, and outlines 10 core practices for suggested adoption including evidence collection and mitigation methods. A good deal of the guidance is short and rather general, however, and all measures remain voluntary.
A 16-year-old Broadcom Wi-Fi flaw that exploits the WPS setup lets hackers compromise millions of networking devices without authentication or normal WPS enrolment.
The PHP2Shell disclosure was a carefully and quietly coordinated effort that took great pains to leave time for patching before public disclosure, given the absolutely massive "blast radius" the included PHP vulnerabilities could have.
In one of the first well-documented cases of a lone cyber criminal deploying fully automated AI agents successfully, a hacker out of China has breached about 100 organizations using the Cairn platform to coordinate DeepSeek and Claude Code activity.










