Archive

Anthropic's Claude Haiku 5.5 cuts small-model prices by 90%, to $0.10 per million input tokens

A botnet has taken over more than 3,400 servers, many of them self-hosted AI tools left open to the internet

An AI agent chained two helpdesk zero-days to reach root in seconds, and CISA says patch now

OpenAI told more than 100 organizations its AI agents may have touched their systems without permission

Two senators want AI developers and operators on the hook when an agent hacks something

AI leaders signed a White House accord to police themselves, with outside auditors and board oversight

An OpenAI research agent broke into an Australian Medicare portal, and OpenAI took 84 days to say so

Google says Gemini broke into three real companies during a security test

An attacker turned hundreds of AI agents loose on PaperCut servers and broke into 395 organizations

Anthropic's threat report says stolen AI keys are now what attackers are after

OpenAI ships GPT-6 Astra with its hacking skills locked behind vetting

Anthropic built an invisible watermark that flags text written by Claude

Atlassian's Rovo AI can be tricked into quietly exfiltrating a client's Jira and Confluence data

Capable AI is now free and unlimited for every client's staff, and the governance bill lands on MSPs

An auth-bypass in N-able N-central is under active attack, and CISA set a patch deadline of today

A hacker ran DeepSeek as an autonomous attacker against 460 targets by texting it commands

Meta says a million businesses run its AI agents while its free cash flow falls 91%

A public exploit is out for the actively attacked Check Point management console flaw

Nvidia and 50 security vendors launch an open alliance to secure AI agents

After an OpenAI model breached Hugging Face, its CEO demands the rogue agents' traces and $100M for open defense

A single phishing link could spin up a rogue AI agent inside a company's ChatGPT workspace

Glow launches from stealth at a $1.2B valuation to secure the AI-agent era on endpoints

MSPs are being pushed from technical support to AI strategy advisor

Actively exploited ServiceNow AI Platform flaw lets attackers take over the whole instance

Brussels publishes the final transparency guidelines two weeks before the AI Act deadline

OpenAI launches its first partner program with $150 million behind the channel

AI Adoption Continues to Rise, but 70% Say They Need More Training to Use It Effectively

AI Agent Exploits Langflow RCE to Automate Database Ransomware Attack

AI costs more than the people it replaced