While the shortest distance between two points is a straight line, a straight-line attack on a large language model isn't always the most efficient — and least noisy — way to get the LLM to do bad ...
Microsoft has pushed back against claims that multiple prompt injection and sandbox-related issues raised by a security ...
A now patched flaw in Microsoft 365 Copilot let attackers turn its diagram tool, Mermaid, into a data exfiltration channel–fetching and encoding emails through hidden instructions in Office documents.
Security researchers have discovered a new indirect prompt injection vulnerability that tricks AI browsers into performing malicious actions. Cato Networks claimed that “HashJack” is the first ...
Network defenders must start treating AI integrations as active threat surfaces, experts have warned after revealing three new vulnerabilities in Google Gemini. Tenable dubbed its latest discovery the ...
New findings from a group of researchers at the Black Hat hacker conference in Las Vegas has revealed that it only takes one "poisoned" document to gain access to private data using ChatGPT that has ...
Feature With great power comes great vulnerability. Several new AI browsers, including OpenAI's Atlas, offer the ability to take actions on the user's behalf, such as opening web pages or even ...
Microsoft has launched Prompt Shields, a new security feature now generally available, aimed at safeguarding applications powered by Foundation Models (large language models) for its Azure OpenAI ...
OpenAI has deployed a new automated security testing system for ChatGPT Atlas, but has also conceded that prompt injection remains an "unsolved" security threat.
OpenAI says prompt injection attacks remain an unsolved and enduring security risk for AI agents operating on the open web, even as the company rolls out new ...
Computer scientists based in South Korea have devised what they describe as an "AI Kill Switch" to prevent AI agents from carrying out malicious data scraping.… Unlike network-based defenses that ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback