Over 175,000 publicly exposed Ollama AI servers discovered worldwide - so fix now
Summary
TechRadar Pro reports that roughly 175,000 Ollama AI deployments are publicly exposed due to misconfiguration, potentially enabling LLMjacking where attackers can generate spam, malware content, or run code remotely. The issue stems from exposing Ollama to the internet rather than binding to localhost; Ollama defaults to localhost, so the risk is caused by user misconfiguration. The article emphasizes fixing exposure by binding to localhost, implementing proper access controls, and treating this as a configuration issue rather than a software bug.