ChatGPT-4o Exploited for Autonomous Voice-Based Scams

Security researchers have demonstrated how ChatGPT-4o, OpenAI's latest voice-enabled AI model, can be abused for autonomous scams. These scams range from bank transfers to credential theft, with success rates reaching up to 60%. Using simple jailbreak techniques, attackers can bypass protections to perform complex scams at minimal cost, highlighting the urgent need for enhanced safeguards.

Read more about this development on BleepingComputer here.