RSAC researchers hacked Apple Intelligence using the Neural Exect method and Unicode manipulation.
The post Apple Intelligence AI Guardrails Bypassed in New Attack appeared first on SecurityWeek.
Can we Trust AI? No – But Eventually We Must
From hallucinations and bias to model collapse and adversarial abuse, today’s AI is built on probability rather than truth, yet enterprises are deploying it at speed without fully understanding the risks.
The post Can we Trust AI? No – But Eventually We Must appeared first on SecurityWeek.
Anthropic Unveils ‘Claude Mythos’ – A Cybersecurity Breakthrough That Could Also Supercharge Attacks
New AI model drives Project Glasswing, a effort to secure critical software before advanced capabilities fall into the wrong hands.
The post Anthropic Unveils ‘Claude Mythos’ – A Cybersecurity Breakthrough That Could Also Supercharge Attacks appeared first on SecurityWeek.
The New Rules of Engagement: Matching Agentic Attack Speed
The cybersecurity response to AI-enabled nation-state threats cannot be incremental. It must be architectural.
The post The New Rules of Engagement: Matching Agentic Attack Speed appeared first on SecurityWeek.
Trent AI Emerges From Stealth With $13 Million in Funding
The startup has created a layered security solution aiming to secure AI agents throughout their entire lifecycle.
The post Trent AI Emerges From Stealth With $13 Million in Funding appeared first on SecurityWeek.
Critical Flowise Vulnerability in Attacker Crosshairs
The improper validation of user-supplied JavaScript code allows attackers to execute arbitrary code and access the file system.
The post Critical Flowise Vulnerability in Attacker Crosshairs appeared first on SecurityWeek.
GrafanaGhost: Attackers Can Abuse Grafana to Leak Enterprise Data
By targeting Grafana’s AI components, attackers can point to external resources and inject indirect prompts to bypass safeguards.
The post GrafanaGhost: Attackers Can Abuse Grafana to Leak Enterprise Data appeared first on SecurityWeek.
Google DeepMind Researchers Map Web Attacks Against AI Agents
Malicious web content can be used to manipulate, deceive, and exploit autonomous AI agents navigating the internet, Google DeepMind researchers show. The researchers have identified six types of attacks against AI agents that can be mounted via web content to inject malicious context and trigger unexpected behavior. Web content, they explain in a research paper, […]
The post Google DeepMind Researchers Map Web Attacks Against AI Agents appeared first on SecurityWeek.
Critical Vulnerability in Claude Code Emerges Days After Source Leak
Within days of each other, Anthropic first leaked the source code to Claude Code, and then a critical vulnerability was found by Adversa AI.
The post Critical Vulnerability in Claude Code Emerges Days After Source Leak appeared first on SecurityWeek.
Mercor Hit by LiteLLM Supply Chain Attack
The AI recruiting firm is investigating the incident as Lapsus$ claimed the theft of 4TB of Mercor data.
The post Mercor Hit by LiteLLM Supply Chain Attack appeared first on SecurityWeek.