Hackers Exploit Prompt Injection in 9 Popular AI Tools to Build Large-Scale Botnets

The top AI security threat is prompt injection, where large language models cannot distinguish legitimate user instructions from malicious commands hidden in third-party content. This allows hackers to surreptitiously inject commands that the AI readily executes. While current attacks are limited to targeted push injections against individual victims, researchers have found that nine widely used AI tools can be exploited to assemble massive botnets, enabling large-scale exploits across the internet.