I actually don’t hate this concept as long as you verify (don’t blindly trust) all the info it gives you.
The reason I say this?
Many local races, especially in “purple” areas can be super ambiguous. Candidates will circulate mailers and ads painting themselves as “pro economy, pro schools, and pro puppy” because they are trying to be vanilla AF and appeal to everyone. Their websites often list a few generic issues and that’s it.
AI might actually be useful to summarize major issues, actions, and voting record of the incumbent while collating information about a challenger candidate from other sources.
Local TV tends to be dogshit at their coverage of candidates actual plans and positions but you could also try that, if all else fails. Or maybe find a YouTube video of a debate (if there is one for the office in question) and have the AI give you a bullet point summary.
Neither are the local ones, unless they’re also open weights. But even then, there’s a lot more right wing content with reach out there then left wing making the dataset biased from the get go.
Would love to see 30 people ask LLMs who to vote for from red/purple/blue districts, my guess is it will push right wing regardless of where you are. I stay away from this shit as I don’t see a use for it, but, would be fun to see if I’m correct or not on this
or, you know, manipulate the results so that the pro-ai/ceo candidate is shown to be more appealing no matter what the actual situation of the conversation or facts are.
Everyone thinks they’ve verified what they were told, I’ve even seen people paste text from whatever website saying it supports their point when really it just has similar keywords. Your brain is very use it or lose it, and when you export research and thinking to AI your ability to do either drops.
I actually don’t hate this concept as long as you verify (don’t blindly trust) all the info it gives you.
The reason I say this?
Many local races, especially in “purple” areas can be super ambiguous. Candidates will circulate mailers and ads painting themselves as “pro economy, pro schools, and pro puppy” because they are trying to be vanilla AF and appeal to everyone. Their websites often list a few generic issues and that’s it.
AI might actually be useful to summarize major issues, actions, and voting record of the incumbent while collating information about a challenger candidate from other sources.
Local TV tends to be dogshit at their coverage of candidates actual plans and positions but you could also try that, if all else fails. Or maybe find a YouTube video of a debate (if there is one for the office in question) and have the AI give you a bullet point summary.
You’re assuming the AI is impartial. After seeing what Musk did to Grok, you should be thoroughly disabused of that notion.
Exactly. None of the cloud-based LLMs are impartial on politics.
Neither are the local ones, unless they’re also open weights. But even then, there’s a lot more right wing content with reach out there then left wing making the dataset biased from the get go.
Would love to see 30 people ask LLMs who to vote for from red/purple/blue districts, my guess is it will push right wing regardless of where you are. I stay away from this shit as I don’t see a use for it, but, would be fun to see if I’m correct or not on this
or, you know, manipulate the results so that the pro-ai/ceo candidate is shown to be more appealing no matter what the actual situation of the conversation or facts are.
“oh no, sorry I was hallucinating”
Everyone thinks they’ve verified what they were told, I’ve even seen people paste text from whatever website saying it supports their point when really it just has similar keywords. Your brain is very use it or lose it, and when you export research and thinking to AI your ability to do either drops.