Former Mt. Gox CEO Mark Karpeles recently submitted the exchange's 2011 codebase to Claude AI, which identified significant security vulnerabilities. The AI analysis labeled the code as "critically insecure," highlighting potential flaws that may have contributed to the infamous collapse of the exchange. This revelation raises questions about whether advanced AI tools could have prevented the security breaches that led to Mt. Gox's downfall.