I don’t really agree with filtering the LLMs but your view is a strawman here.
The risk isn’t the content, it’s the scale at which the content can be produced and personalized.
For example, this makes it extremely easy for a well funded extremist group to start interacting with people on social media through an LLM slowly convincing them that we should imprison all ${something}. Over a few years you’ll potentially see that movement become legitimized and mainstream.
That risk isn’t something you’re going to be able to mitigate by hoping every LLM is filtered. But it is a risk our society now needs to grapple with though.
The risk isn’t the content, it’s the scale at which the content can be produced and personalized.
For example, this makes it extremely easy for a well funded extremist group to start interacting with people on social media through an LLM slowly convincing them that we should imprison all ${something}. Over a few years you’ll potentially see that movement become legitimized and mainstream.
That risk isn’t something you’re going to be able to mitigate by hoping every LLM is filtered. But it is a risk our society now needs to grapple with though.