The release of two malicious language models — WormGPT and FraudGPT — demonstrate attackers' evolving capability to harness language models for criminal activities.
See All 3141 Episodes of "The Latest"