'Inner Thoughts' of Every Major AI Model Exposed in Massive Exploit
60-second summary
Researchers have exposed the inner workings of every major AI model by exploiting a global key used to encrypt reasoning tokens. They decoded 315,320 hidden thinking blocks from public logs, recovering sensitive information like passwords and live API keys in the process.
Researchers found every major AI provider encrypts reasoning tokens with a single global key—and exploited it to decode 315,320 hidden thinking blocks from public logs, recovering passwords and live API keys along the way.