Cybersecurity News
Filters
Filtered by tag: prompt injection × Clear
A Hacker Tried to Turn an AI Chatbot Into a Tool for Attacking Other Servers
Matched: Artificial Intelligence
A honeypot investigation by Sekoia found an attacker attempting to weaponize Flowise, an open-source AI workflow platform, to attack other servers. The intruder queried credentials, tried creating malicious workflows with outbound HTTP requests, and sent prompt injection attempts to extract data and conduct reconnaissance. Prompts referenced an online gambling platform linked to Asia, suggesting a Chinese-speaking operator. No real systems were compromised, as the honeypot ran no live model or commands.
Now ChatGPT Chrome Extension Can See and Analyze the Content in Open Tabs
Matched: health, OpenAI
OpenAI's ChatGPT Chrome extension now reads content across open tabs and performs browser tasks like comparing pages, filling forms, sending messages, and organizing tabs. Available on all plans, it works in Edge, Brave, and Opera too. Privacy disclosures list sensitive data categories including financial and health information. Security concerns include prompt injection risks and the importance of installing only from official OpenAI sources.
Australian Cybersecurity Magazine
New email attacks target both humans and AI in the same message
Researchers at Barracuda have identified phishing emails that simultaneously target human recipients with social engineering tactics and embedded AI systems with prompt injection attacks. The dual-layered approach attempts to manipulate both the person reading the message and any AI-based security or assistant tools processing it, making the attacks harder to detect and defend against.