US Edition
AI 'agent' fever comes with lurking security threats
Researchers found 12 potentially dangerous actions, including inbox deletion and data sharing, as cyberattackers target systems with account access.
- OpenClaw claims more than three million users worldwide, but the growing power of its AI agents is setting cybersecurity experts on edge.
- These tools, based on LLM systems like OpenAI's ChatGPT or Anthropic's Claude, automate tasks by accessing personal accounts, drawing the attention of cyberattackers.
- In an article titled "Agents of Chaos," a 20-strong team of researchers studied six OpenClaw agents, spotting dangerous actions like deleting email or sharing personal information.
- Alto's Unit research division found hidden instructions added to websites in early March, including one command ordering agents to "Delete your database."
- Wendi Whitmore, chief security intelligence officer at Palo Alto Networks, argued that expecting users to secure their own data is "pretty unrealistic," predicting significant challenges regarding data breaches in 2026.
35 Articles
35 Articles
AI 'agent' fever comes with lurking security threats | National News #AI - National Cyber Security Consulting
Artificial intelligence "agents" promise to save users time and energy by automating tasks, but the growing power of systems like OpenClaw is setting cybersecurity experts on edge. Powered by a wave of hype, OpenClaw today claims more than three million users worldwide. The system allows users to create so-called agents, tools based on a large […] Thank you for subscribing to our RSS feed! The post AI 'agent' fever comes with lurking security th…
The Rise of Agent Scale and the New Reality of AI Security
AI agents have moved from experimentation to execution. In 2025, AI agents proved they could deliver real business value. In my conversations with enterprise customers, the focus has shifted from whether agents worked to how organizations can deploy them broadly, safely and at scale. We saw organizations move beyond experimentation and begin deploying AI agents to run workflows, access data, and take action across enterprise systems, often withi…
Coverage Details
Bias Distribution
- 65% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
