I don’t hate AI and think it’s fine to use for a bunch of things, but using it to falsify the level of public engagement on a political issue is a clear misuse, it’s easy to see how that could make democracy not work as well, or backfire and be used as an argument that all the public sentiment about the issue is astroturfed.
Well, the person you responded to above was talking about sending more than one, which is the worst part. But even if you are only using AI to rephrase the canned response for your singular comment, that creates a situation where it is more difficult for them to actually read and consider different points people might be bringing up, because now there’s lots of messages that are basically just the canned response in content and intent but more effort to group together. Also the people going through them will probably be able to tell AI is being used, which could call into question whether someone was sending more than one even if you were not.
It’s possible, but I’ve followed some public comment processes for regulatory stuff before and large volumes of comments make it take way longer, because there is manual work involved. If a politician wants to still have actual people manually consider the contents of their inbox (which they absolutely should), using AI instead of a form letter will make that much harder for them to do. AI talking to AI to determine what the public thinks and wants is probably going to lose a lot in translation, and if it’s using service-based AI will give the companies running it another rather direct way to influence political outcomes.
Given all that, I’m not sure what the advantage is to balance against it either. As opposed to sending a copy of the form letter, where you can assume they will at least count how many people have done that, what’s even the benefit of having a LLM rewrite it first?
I don’t hate AI and think it’s fine to use for a bunch of things, but using it to falsify the level of public engagement on a political issue is a clear misuse, it’s easy to see how that could make democracy not work as well, or backfire and be used as an argument that all the public sentiment about the issue is astroturfed.
Is it any worse than using the canned response from the page? I’m obviously interested, I’m just using ai to make it easier.
However a hand written note is probably the most efficient.
Well, the person you responded to above was talking about sending more than one, which is the worst part. But even if you are only using AI to rephrase the canned response for your singular comment, that creates a situation where it is more difficult for them to actually read and consider different points people might be bringing up, because now there’s lots of messages that are basically just the canned response in content and intent but more effort to group together. Also the people going through them will probably be able to tell AI is being used, which could call into question whether someone was sending more than one even if you were not.
They’re not reading the emails, they probably have an ai to combine and summarise. So then that ai can talk to my ai. 😁
It’s possible, but I’ve followed some public comment processes for regulatory stuff before and large volumes of comments make it take way longer, because there is manual work involved. If a politician wants to still have actual people manually consider the contents of their inbox (which they absolutely should), using AI instead of a form letter will make that much harder for them to do. AI talking to AI to determine what the public thinks and wants is probably going to lose a lot in translation, and if it’s using service-based AI will give the companies running it another rather direct way to influence political outcomes.
Given all that, I’m not sure what the advantage is to balance against it either. As opposed to sending a copy of the form letter, where you can assume they will at least count how many people have done that, what’s even the benefit of having a LLM rewrite it first?