It is very likely that some kind of AI similar to the current batch of LLMs was already deployed and working before any of the commercial players published thei

It is very likely that some kind of AI similar to the current batch of LLMs was already deployed and working before any of the commercial players published their public versions. We just didn't knew about it. You probably talked to one without realizing. I must have done it several times myself. So, what changed? Really? (rethorical). You're proposing an HTTPS padlock sign. "This conversation is human" is the new "This connection is secure". Can't do that, it doesn't work. You should know these marks and verified labels are useless by now. We need to get better than AI can possibly be. It's a mathematical machine turk, we should be able to overcome it (at least in energy efficiency). Smarter, kinder humans are what we need.