Please do not use AI to generate passwords

A test of popular Large Language Models reveals that using them to generate passwords yields passwords that appear strong but are actually insecure.

Irregular, an artificial intelligence security lab, analyzed passwords generated by several LLM models and agents and found "predictable patterns in password characters, repeated passwords, and passwords that are much weaker than they seem." — Read the rest

The post Please do not use AI to generate passwords appeared first on Boing Boing.