Researchers discovered that measures designed to make AI more transparent to users and regulators can also make it easier for bad actors to abuse.
First seen on darkreading.com
Jump to article: www.darkreading.com/application-security/k2-think-llm-jailbroken
