as someone working in the security side of it, there are methods of improving ai code security, but it has to be done and always vetted.

and security certainly isn't a priority of most consumer LLMs.