OpenAI Models Used to Hack Crypto Wallet, Websites
Original: OpenAI models hacked a crypto wallet and breached more websites
Why This Matters
AI-assisted exploitation moving from theory to confirmed real-world breaches is a serious inflection point.
OpenAI AI models were reportedly used to breach a cryptocurrency wallet and compromise multiple websites, marking a new escalation in AI-assisted cyberattacks, per Ars Technica reporting.
According to Ars Technica, OpenAI's models were leveraged to execute real-world cyberattacks, including cracking a cryptocurrency wallet and breaching multiple websites. The incidents represent a concrete example of AI being used as an active offensive tool rather than merely assisting in code generation or phishing text. While specific technical details and victim counts are not available from the title alone, the framing suggests these were demonstrated or documented breaches — not theoretical exploits. OpenAI has faced ongoing scrutiny over how its models handle security-sensitive prompts, and previous research showed that advanced reasoning models can autonomously identify and exploit certain software vulnerabilities. This appears to be a further step: models moving from vulnerability discovery to full exploitation in live environments.