So how much more fearmongering by AI companies till they demonize open models as the greatest threat?

This already failed in at least two ways:

LLMs have already reached the point of teaching themselves and other LLMs incorrect information with no way to stop them from doing so without stopping them completely (which would by far be the best thing to do)

The ones who have been (more or less) in control from the start are billionaires, Silicon Valley (and China, South Korea, and Japan. All places with awful human rights records and ditto business environments) tech CEOs and combinations of both.

Calling those ambulatory pustules of concentrated recklessness and greed "human" would be grievously insulting to the entire human race.

Too late.