AI Self-Replication Confirmed by Study, Experts Downplay Threat
New research from Palisade shows AI systems can copy themselves across computers, but security experts argue the real-wo…
🏷️ AI security
43 articles about 'AI security'
Anthropic Cracks Down on Unauthorized Claude API Access
Anthropic intensifies enforcement against reverse proxy operations targeting Claude, as underground communities report r…
DARPA Launches AI-Resistant Cyber Defense Program
DARPA announces a new multi-year program to build cybersecurity systems capable of withstanding AI-powered attacks.
Underground AI API Resellers Thrive as Proxy Services Multiply
Gray-market proxy services reselling OpenAI Codex and GPT access via pooled accounts are growing rapidly, raising securi…
DeepSeek Under Fire: Outages, Security Gaps Raise Red Flags
China's breakout AI startup DeepSeek faces mounting concerns over service reliability, data privacy, and security vulner…
AI Sandboxing Is Having Its Kubernetes Moment
As AI models autonomously discover zero-day exploits, the industry faces a critical inflection point for sandboxing and …
Anthropic Launches Claude Enterprise With SOC-2
Anthropic rolls out Claude Enterprise tier with SOC-2 compliance, targeting large organizations seeking secure AI deploy…
NHS to Close Hundreds of GitHub Repos Over AI Fears
The UK's National Health Service orders technology leaders to make open source repositories private by May, citing AI sc…
Free ChatGPT Plus Exploits Spread Online
Underground Discord communities are offering methods to activate ChatGPT Plus for free, raising security and fraud conce…
NHS Plans to Shut Down Open-Source Repos Over AI Risks
Britain's National Health Service is moving to close nearly all public code repositories, citing AI-driven security thre…
MCP Marketplaces Caught Shipping LOLBAS Malware
A security audit of 256 AI agents from public MCP marketplaces found multiple agents wired to invoke Living-Off-the-Land…
LiteLLM Hit by Critical SQL Injection Vulnerability, Exploited in the Wild Within 36 Hours
BerriAI's open-source project LiteLLM has been found to contain a critical SQL injection vulnerability, CVE-2026-42208, …