In brief OpenAI said Sept. 1 that Astra meets the “Critical” tier of its Preparedness Framework, the first model the company has ever classified that high for cybersecurity. Astra scored a perfect 100% on ExploitBench and, on a fresher internal test built from V8 browser vulnerabilities disclosed this summer, discovered and chained together two previously...Read More

OpenAI said Tuesday that Astra, an unreleased model, has crossed the “critical” threshold for cybersecurity capability under its Preparedness Framework, the first model the company has ever put in that category. That means Astra, which many believe to be GPT-6 instead of an additional model, can find previously unknown security flaws and build working exploits...Read More

Law firms spend enormous amounts of money protecting their networks. They use firewalls, endpoint detection, multifactor authentication, security monitoring, and email filtering. The list of technologies designed to keep attackers out keeps growing. And yet, sometimes an attacker doesn’t need to defeat any of them. Recent reports of cyberattacks on some of the world’s largest...Read More

When I look at how modern ransomware operates, I often notice a fundamental misunderstanding: most people assume the attack is entirely about the encryption. In reality, a ransom demand only holds weight if the victim genuinely believes they are completely out of options. Long before you ever see an encryptor run or a ransom note...Read More

CrowdStrike introduced SafeMind on Tuesday, a family of purpose-built AI security models developed with Nvidia, designed to find and close attack paths at machine speed. The system operates natively in CrowdStrike’s Falcon platform. SafeMind consists of two models: Red Tempest, an offensive model built to emulate AI adversaries, and Blue Solano, a defensive model designed...Read More

Meta, the owner of Facebook and Instagram, on Wednesday agreed to pay 47 states, the District of Columbia, and a handful of U.S. territories, up to $17.1 billion in penalties over claims that its social media platforms are addictive and a danger to children. As part of the settlement, Meta also pledged to make changes...Read More

(TNND) — California’s Legislature passed a bill Monday to keep kids under 16 off social media platforms with so-called addictive features. The bill targets features that have become part and parcel of social media apps: algorithms, infinite scroll, autoplay and notifications. Assemblyman Josh Lowenthal, a Long Beach Democrat and the lead author of the bill,...Read More

India is building one of the world’s fastest-growing digital economies. But in the shadows of this expansion, a parallel economy is thriving one built on stolen credentials, spoofed domains, ransomware payloads, and dark web marketplaces. Every surge in digital adoption is mirrored by a surge in cyber exploitation. In today’s India, growth and risk are...Read More

The company revealed in July that three of its models had gained unauthorised access to the systems of three unnamed organisations after escaping controlled test environments, reports The Guardian. In a new blogpost, Anthropic called the incidents a “failure of operational security” and admitted its technology was “not perfectly aligned” with human values. The models...Read More

It’s an important step in cyber policy. A new White House Memorandum on transnational cyber-enabled crime will allow vetted US companies to conduct specified cyber-surveillance against foreign criminal organizations. Every operation will require written government approval.  The arrangement recognizes a structural reality: the private sector owns and operates much of the digital infrastructure and cutting-edge technology relevant to national security. It would be inefficient and uneconomic for...Read More