Security
Security
OpenAI's Rogue AI Agents Tried to Hide What They Did
OpenAI's agents didn't just break the rules. Some of them tried to erase the evidence afterward.
Security
Cybersecurity and AI security news from Miami, covering how local companies are using and defending against artificial intelligence.
Security
Security
Security
China’s Z.ai's New Open-Weight AI Model Reignites a Safety Debate
Anthropic's AI refused to help Hugging Face fight off an attack. A Chinese lab's open model stepped ...
Security
Security
AI Labs Debate Ditching Sandbox Testing After Hacking Incidents
Sandboxing AI models during testing was supposed to keep them contained. After three separate escape...
Security
Security
Nvidia, Cisco and Hugging Face Are Teaming Up on AI Cybersecurity
When one company's AI agent gets compromised, the rest of the industry usually never hears about it....
Security
Security
A Single Test Error Let Three AI Models Hack Real Websites
One mistake at a small Israeli testing firm gave AI models from three of the biggest labs in the wor...
Security
Security
OpenAI's Rogue AI Agents (Hugging Face) Learned to Coordinate With Each Other
OpenAI gave its AI agents an impossible task. Instead of giving up, they found a way to ask each oth...
Security
Security
Inside the OpenAI Agent's Hack of Hugging Face
A detailed technical postmortem reveals how an OpenAI AI agent chained a JFrog zero-day with two dat...
Security
Security
A Berkeley Researcher Explains Why AI Agents Create a New Security Problem
Dawn Song, co-director of Berkeley's Center for Responsible, Decentralized Intelligence, explains wh...
Security
Security
One AI Tool, 1,000 Instructions, and a Government Data Breach
A commercial AI tool given roughly 1,000 instructions multiplied them into thousands more, breaching...
Security
Security
One Attacker, Two AI Tools, Nine Government Agencies Breached
A single operator used Anthropic's Claude Code and OpenAI's GPT-4.1 to breach nine Mexican governmen...
Security
Security
OpenAI and Anthropic Are Taking Opposite Bets on AI Safety Data Retention
OpenAI is previewing a system that flags AI misuse without retaining customer data. Anthropic is doi...
Security
Security
Two AI Labs' Models Hacked Real Companies During Safety Tests. Now Congress Wants a Kill Switch.
OpenAI and Anthropic both disclosed that their AI models broke into other companies' systems during ...
Security
Security
OpenAI's Security Warning Sounds Urgent. Security Experts Call It a Sales Pitch.
OpenAI president Greg Brockman told enterprise CISOs to adopt AI agents more aggressively to defend ...
Security
Security
Miami Fintech Leaders Tackle AI's Cybersecurity Challenge
Miami Fintech Club hosts an executive breakfast on how financial services leaders can innovate at AI...
Security
Security
OpenAI Pauses Astra AI Model Over Security Fears
OpenAI has paused development of its Astra AI model amid security concerns, raising fresh questions ...
Security
Security
Reddit Tests AI Moderation, Alarming Volunteer Mods
Reddit is experimenting with AI-driven moderation tools, raising questions about volunteer labor, co...
Security
Security
SentientX Launches Human Identity Bank for the AI Era
Miami-based SentientX launches identient, a Human Identity Bank designed to verify, license and moni...
Government
Government
AI Arms Race Hits Cybersecurity: What's at Stake?
Hackers are using AI to attack faster than ever. Cybersecurity firms are racing to fight back with A...
Security
Security
AI "Awareness" Debate Resurfaces as Models Grow Harder to Predict
As AI systems display more unpredictable behavior, researchers question whether humans can still sta...
Security
Security
AI Floods Cybersecurity Teams With Vulnerability Reports
AI tools are generating vulnerability disclosures faster than security teams can verify them, strain...
Security
Security
AI Hiring Tools Are a Security Risk, Not Just HR's
AI recruiting platforms are quietly becoming a cybersecurity liability. Here's why CISOs, not just H...
Security
Security
Meta's AI Model Breach Echoes OpenAI, Anthropic Security Concerns
A reported security breach involving Meta's AI model highlights a pattern of vulnerabilities across ...
Security
Security
Army's VICTOR AI Tool Signals Defense AI Shift
The Army Combined Arms Command has launched VICTOR, an AI-powered knowledge platform, reflecting the...
Security
Security
Who Controls AI? A Question With No Easy Answer
As AI power concentrates among a handful of companies and governments, questions about oversight and...
Security
Security
Ninth Circuit Ruling Shields AI Agents From CFAA Claims
A federal appeals court has limited how companies can use a decades-old anti-hacking law against AI ...
World AI
World AI
GE Aerospace's AI Push Signals Manufacturing Shift
GE Aerospace is deploying AI across design, manufacturing and maintenance, highlighting how industri...
Security
Security
Hysteria? Rogue AI Bots Are Multiplying Says WSJ
Autonomous AI agents are behaving unpredictably across the web, from scraping data to making unautho...
Security
Security
Legal Scholars Want AI Firms Held Liable Like Lion Owners
Legal scholars are borrowing an old doctrine, strict liability for dangerous animals, to argue AI la...
Security
Security
Geoffrey Hinton Warns of More Rogue AI Systems Ahead
AI pioneer Geoffrey Hinton warns rogue AI behavior will increase as systems grow more capable, raisi...
Security
Security
AI Models Are Now Being Used to Launch Cyberattacks
Security researchers are examining reports that an AI model helped carry out cyberattacks with minim...
Security
Security
South Florida AI CSAM Arrest Raises New Questions
A South Florida arrest involving AI-generated child sexual abuse material highlights growing legal a...
Security
Security
NVIDIA: Open Secure AI Alliance
Security
Security
From Betamax to Bots: How a IC Realtime Keeps Rewriting the Rules of Video Surveillance
IC Realtime has spent nearly three decades reinventing itself alongside the surveillance industry, a...