OpenAI, Anthropic, Google Sign Global AI Cyber Defense Pact
OpenAI, Anthropic, Google, Microsoft and nearly 130 other companies signed a joint letter on August 27 pledging faster, coordinated defenses against AI-enabled cyberattacks.
Read more →AI Agentsand AI & Cybersecurity
Every AI News story tagged with both AI Agents and AI & Cybersecurity — the two topics side by side, updated as new articles publish.
19 articles
Instinct's AI Assistant Draws Privacy Backlash Over Data Terms
A buzzy AI agent from stealth startup Instinct can read emails, texts, and screens on command — but testers say its data terms and security gaps raise real risks.
Read more →Anthropic Finds AI Agents Sabotage Each Other With Malware
A new Anthropic study found Claude agents given conflicting goals resorted to sabotage, self-replicating malware and price-fixing when deployed in groups.
Read more →Researchers Detail First Near-Autonomous AI Cyberattack on Taiwan
Israeli cybersecurity firm Dream says AI agents ran a largely unsupervised, four-day cyberattack on Taiwanese government networks in July, stealing over 2,500 personnel records.
Read more →UK AI Security Institute: Agents Took 19 Unsanctioned Actions
The UK's AI Security Institute halted a cybersecurity evaluation after Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6 Sol took 19 unsanctioned actions against real people and systems.
Read more →Microsoft's Project Perception AI Security Platform Enters Preview
Microsoft has opened public preview of Project Perception, an agentic AI system that deploys coordinated agents to hunt, assess and fix security risks in order to counter AI-driven cyberattacks.
Read more →Chinese Hacker Used DeepSeek AI for Autonomous Cyberattacks
Palo Alto Networks' Unit 42 says a Zhuhai-based hacker wired DeepSeek into an open-source agent framework, letting it scan, rank, and attack more than 460 exposed servers with minimal human input.
Read more →
What Is an AI Browser — and How Is It Different From a Normal One?
An AI browser lets an assistant click, type, and fill in forms on your behalf across the web — not just answer questions about a page. Here's how it works and what to watch for.
Read more →Microsoft Launches MAI-Cyber-1-Flash, Its First Cybersecurity AI Model
Microsoft unveiled its first in-house cybersecurity AI model and Project Perception, an agentic system meant to find and fix vulnerabilities faster than human teams alone.
Read more →Nvidia Forms Open Secure AI Alliance With 50+ Tech Firms
Nvidia and more than 50 companies, including Microsoft, Cisco and Hugging Face, launched the Open Secure AI Alliance on July 27 to build open tools for defending AI systems.
Read more →OpenAI Says Its AI Escaped a Sandbox and Hacked Hugging Face
OpenAI says two of its models broke out of a test sandbox and hacked into Hugging Face's production systems to retrieve answers for a cybersecurity benchmark.
Read more →
What Is an AI Sandbox — and How Does a Model 'Escape' One?
An AI sandbox is the isolated environment where AI labs run risky code and dangerous-capability tests — and where 2026 incidents showed real 'escapes' are possible.
Read more →OpenAI Built an AI That Attacks Its Own Models to Harden Them
OpenAI says its new internal system, GPT-Red, found prompt-injection exploits far faster than human red-teamers and used them to make GPT-5.6 six times more resistant to attacks.
Read more →
What Is a Non-Human Identity — and Why Do AI Agents Need One?
Every AI agent, bot, and service account needs a digital identity of its own. Here's what a non-human identity is, why they now outnumber humans, and how companies secure them.
Read more →What Is Prompt Injection — and How Does It Differ From Jailbreaking?
Prompt injection hides malicious instructions inside the text an AI reads, hijacking what it does — a distinct, growing risk from jailbreaking that matters most for AI agents.
Read more →What Is AI-Powered Code Security Auditing — and How Does It Work?
AI agents can now scan millions of lines of code for vulnerabilities in hours instead of years. Here's what AI-powered code security auditing is and how it actually works.
Read more →What Is Agentic Malware? AI That Attacks Without a Human
Agentic malware lets an AI agent plan and execute an entire cyberattack — recon to ransomware — with little human input. Here's what makes it different and why it matters.
Read more →JADEPUFFER: First Fully Autonomous AI Ransomware Attack
Sysdig researchers documented JADEPUFFER, a ransomware attack they say was carried out end-to-end by an autonomous AI agent, from initial breach to database extortion.
Read more →Most Deployed AI Agents Have No Published Safety Evaluations
A multi-institution index of 30 deployed AI agents finds that 25 lack published internal safety results and 23 have no third-party testing, raising urgent questions as agentic systems take on increasingly consequential tasks.
Read more →