Raw LLM Responses

Re generated images, it all sounds too easy to flag them as such, but any modern phone uses image processing to produce much-wanted enhancements as a matter of course, so there is a slippery slope between something that is not considered 'generated' and something that is, and if we were accurate, we would flag ALL images produced by phone cameras as 'generated', which then defeats the idea of segregating generated and 'non-generated' images, unless we introduce some sort of 'levels of generatedness' or 'levels of unrealness' or 'levels of deceit' or 'levels of imagination or creativity'. Even looking at karaoke machines that have been around for years, some of them automatically adjust the pitch of a singer's voice so that it is in tune, and that is another example of a lie that we ignore, but again it is on the edge of a slippery slope. So many YouTube videos use stock footage for illustration, and news agencies have for years used old videos and photos of politicians and celebrities to 'illustrate' them when they can't access something more recent or more directly related to the topic of a news item. Yet we don't flash the word 'simulation' under these or require a comment to that effect in narratives. So separating 'generated' content from 'non-generated' is a can of worms. In fact 'non-generated' simply does not exist. And 'generated' doesn't require AI to be misleading (though it obviously helps a hell of a lot).