Friday, April 17, 2026
Tag:

data exfiltration

OpenAI Introduces Lockdown Mode to Combat Prompt Attacks

OpenAI has launched Lockdown Mode, an optional security feature in ChatGPT designed to enhance user safety. This mode focuses on mitigating prompt injection attacks,...

‘ZombieAgent’ Vulnerability Reveals Critical Blind Spots in ChatGPT AI

Cybersecurity firm Radware recently identified a critical vulnerability known as “ZombieAgent,” which exploits OpenAI’s ChatGPT-based AI agents. This zero-click indirect prompt injection flaw enables...

Backslash Launches MCP Security to Protect AI Coding

Backslash Security has introduced a new MCP Security product designed to safeguard Model Context Protocol (MCP) servers amid the prevalent use of AI coding...

BeyondTrust CTO Sounds Alarm on AI-Driven Identity Threats

In today’s cybersecurity landscape, identity has become the new perimeter, asserts James Maude, FCTO at BeyondTrust. Modern attackers exploit stolen or misconfigured identities to...

Gemini Security Vulnerabilities Put Millions at Risk of Undetected Data Breaches

Tenable recently uncovered three critical vulnerabilities in Google’s Gemini suite, termed the “Gemini Trifecta.” These flaws, now patched by Google, jeopardized millions of users...

Navigating New Risks and Urgent Security Demands Posed by AI Agents in Business

Radware's report, "The Internet of Agents: The Next Threat Surface," explores cybersecurity risks associated with the rising use of agentic AI systems, particularly those...

Strategies Used by Attackers to Compromise AI Tools and Defenses

Barracuda highlights the evolving threats of generative AI in crafting spam emails and sophisticated phishing attacks. Recent analyses reveal that attackers manipulate companies' AI...

Potential Hazards of Cloud-Based AI Solutions

As companies increasingly adopt cloud-based AI applications, managing cybersecurity becomes more complex, particularly with the emergence of generative AI and threats like "shadow AI."...