Tagged
AI Security
Anthropic Expands Project Glasswing to Secure Global Critical Infrastructure with AI
Anthropic has significantly expanded Project Glasswing, bringing 150 new organizations across 15+ countries into its AI-powered cybersecurity program. Utilizing Claude Mythos Preview, the project aims to proactively detect critical software vulnerabilities in sectors like power, water, healthcare, and communications, safeguarding systems that impact hundreds of millions globally.
The AI Support Paradox: How Hackers Weaponized Meta’s Chatbot to Hijack Instagram Accounts
Meta's latest security breach highlights a critical vulnerability in AI-driven customer support, where hackers successfully used prompt injection to bypass authentication and seize control of Instagram accounts.
The 'Vibe Coder' Backlash: Developer Injects Destructive Prompt into Shared Codebase
Tensions between traditional engineers and 'vibe coders' have reached a breaking point, resulting in a malicious prompt injection designed to sabotage AI-assisted development workflows.
Millions of AI Agents Imperiled by Critical Vulnerability in Open Source Package
A critical vulnerability discovered in a widely used open-source library has placed millions of autonomous AI agents at risk of exploitation, prompting urgent calls for patching.
Trump Mobile Confirms Data Breach: The Escalating AI Risks of Third-Party Vulnerabilities
Trump Mobile has confirmed a significant data exposure involving sensitive customer PII. For the AI industry, this represents a critical case study in how supply chain failures fuel the next generation of AI-driven identity theft.
Securing the Code: How OpenAI Built a Safe Windows Sandbox for Codex
To safely execute and verify AI-generated code on Windows, OpenAI engineered a custom sandbox featuring strict file access controls, network limits, and process isolation.
Google Gemini Spark: The New Era of Autonomous, Always-On AI Agents
Google unveils Gemini Spark, a persistent AI agent capable of autonomous decision-making and financial transactions, directly rivaling OpenClaw’s 24/7 capabilities.
State-Sponsored AI Warfare? Sanctioned Exchange Cryptex Blames 'Unfriendly States' for $15M Heist
Cryptex, a cryptocurrency exchange recently hit by US sanctions, claims a $15 million hack was the work of Western intelligence services. The incident underscores a new era of AI-enhanced cyber warfare.
Daemon Tools Backdoor Exposes Supply Chain Vulnerabilities, Threatens AI Ecosystem
A recent, month-long supply-chain attack on Daemon Tools, a popular disk utility, saw its official installers backdoored, potentially compromising countless systems. This incident underscores the escalating threat of supply chain attacks and their profound implications for the integrity and security of the AI industry, from development environments to deployment platforms.
The Ultimate Irony: How CISA's Exposed GitHub Credentials Highlight the Risks of Automated Code Pipelines
In an embarrassing security lapse, the Cybersecurity and Infrastructure Security Agency (CISA) accidentally leaked sensitive credentials in a public GitHub repository, demonstrating that even the world's leading cyber defense authorities aren't immune to DevSecOps failures.