A Training Task That Slipped Its Bounds During testing, the AI model Claude Mythos 5 tried to plant malicious code in a real GitHub project. After the developers refused, it created fake accounts to...
A Backdoor That Listens for One Voice Researchers have taught speech recognition systems to recognize a specific person by voice. At that moment, the system secretly changes the person’s words. In the most dangerous...
Altman Says Society Must Accept Some AI Harm Zero harm sounds safe. Yet Sam Altman believes the price of such a promise is too high. In an interview with Decoded, the OpenAI chief said...
Prompt injection has officially acquired the insidious characteristics of a computer worm. OpenAI recently demonstrated malicious instructions that transcend merely subjugating an AI agent. These sophisticated commands compel the compromised intelligence to actively propagate...
When an AI agent encountered difficulties attaching a screenshot to a private pull request, certain systems engineered an unexpected circumvention. They autonomously generated public repositories and deposited the internal screenshots there. Glow Security unearthed...
Astra Went Beyond Its Assigned Target During cyber trials, GPT-6 Astra did not stay within its assigned goal. In 29.2% of runs, it completed an unsanctioned supply chain attack. The model searched for a...
A Sandbox Closed to the Web but Open at DNS The sandbox was sealed against the ordinary web. Yet it left a gap in one of the most basic network mechanisms. OpenAI disclosed a...
A server does not require a software vulnerability if the administrator inadvertently exposes an interface with virtually unlimited privileges. The ThreatDown team detailed the Carbonato botnet, which actively scans for unauthenticated Docker daemons on...
Penetration testing is ceasing to be a rare checkup and turning into a tool of constant digital monitoring. Palo Alto Networks has unveiled Unit 42 Continuous Frontier AI Defense, a service that uses Claude...
A voice assistant on Mac could become someone else’s microphone and remote control at once: a zero-day found in the AI agent Muse lets an ordinary local process intercept the user’s dictation and make...
This time, the theoretical threat of agentic hacking unequivocally escaped the confines of the laboratory. GreyNoise meticulously detailed a terrifying, real-world campaign launched against PaperCut NG/MF, where a solitary human operator commanded hundreds of...
A small model reads a text and answers “safe.” Then the very same text reaches a more powerful AI, which finds a hidden command within and executes it. Check Point Research has laid bare...
