The thought that an LLM hooked up to the web can automatically write callout posts—with all the attention to reality that is to be expected of an LLM—is quite worrying for me, a member of a minority group where individuals are routinely socially murdered by means of callout posts which more often than not contain factual inaccuracies and outright fabrications.
Discussion
Loading...
@vtrlx it's not clear to me that it can, given that every demo of this technology is routinely shown to have been subtly altered or completely faked. if it serves the purpose of propaganda for a highly profitable cause, i believe the burden of proof should always be set much higher than normal
@vtrlx the danger of people accepting that an LLM can generate targeted content without human input is the same danger you describe but slightly rephrased to suit my mental model. the result is that people will not consider the targeted callout post an example of violent human intent and thus fail to disincentivize its creation
@vtrlx seeing the CIA guy lean into "AI" making it impossible to trust any proof of wrongdoing a few weeks ago has not left me. i'm aware this is kind of an unhelpful reply from someone who does not face said danger and i am trying to figure out how to merge the two perspectives
@vtrlx rasha abdulhadi has described the purpose of propaganda being to induce fear and despair
@hipsterelectron I think I understand where you're getting at. For what it's worth, I am not entirely doomering over this idea. I still believe that the proliferation of slop will make honest people more skeptical of what they read both online and in the grass-touching zone. That's not going to stop bias, though.
If anything, I don't know what to think about my idea.