End-to-end encryption is the one of the most basic requirements for a communication system to be secure. Endpoint authentication is another. Message authentication is the third. After those 3 are fulfilled, further requirements can vary from system to system.

It’s like electrical or building code. Just because it’s compliant doesn’t mean it’s safe, but if it’s not compliant it’s almost certainly not safe. Necessary, not sufficient!

The “AI safety” field is about two things: marketing AIs as so powerful that they’re risky to use but riskier to get left behind by competitors using, and keeping AIs from doing so much brand damage that stock price suffers. This story is about marketing an AI as powerful.

LLMs are perfectly suited to make propaganda. That (and advertising, which is just corporate propaganda instead of government propaganda) is their best use. It doesn’t matter that they’re bullshit generators when the goal is to generate bullshit.