#AI Security

Researchers discover ZombieAgent, a persistent vulnerability in ChatGPT that uses long-term memory to steal private data stealthily. Learn more about the ChatGPT ZombieAgent vulnerability.

CrowdStrike acquires identity management startup SGNL for $740 million. Learn how this strategic move aims to defend against AI-driven cyber threats and consolidate the security market.

OpenAI has officially admitted that prompt injection attacks are a permanent, unsolvable threat. A VentureBeat survey reveals a critical gap, with 65% of enterprises lacking dedicated defenses.

ServiceNow announces a $7.75 billion cash deal to acquire cybersecurity startup Armis, aiming to build an AI-powered security platform and capitalize on the growing need to protect enterprises from sophisticated AI threats.

OpenAI concedes that prompt injection, a core AI security flaw, is 'unlikely to ever be fully solved.' We analyze their new defense—an AI-powered attacker—and the expert consensus on the risks of agentic AI.

OpenAI is using automated red teaming with reinforcement learning to strengthen ChatGPT Atlas against prompt injection attacks, creating a proactive loop to discover and patch exploits early.

Google Cloud and Replit executives warn that the hype for 2025 being the "year of the AI agent" is premature. They cite major hurdles in reliability, data integration, enterprise culture, and security.

An analysis of Australia's Bondi Beach attack, exploring its geopolitical implications, impact on global security, and the role of technology in fostering societal resilience.

An AI vending machine lost $1,000 to simple tricks, revealing a critical vulnerability in corporate AI that every business leader needs to understand now.