(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
. . . . . . . . . .
Here's how we should use AIs like Chat GPT [1]
['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.']
Date: 2025-01-28
This post is written by a human with no AI assistance. As it (mostly) should be.
There continues to be a great deal of consternation about the emerging AI chatbots like ChatGPT, Google’s Gemini, and many others. And with good reason. Their “hallucinations” resulting in ridiculous and potentially dangerous falsehoods are numerous. Example articles are easy to find.
That’s because a good deal of data these bots are referencing, the stuff they’re “trained” on, is itself flawed. Garbage in, garbage out.
But that changes when you control the information the bots are sourcing. As this Wapo reporter did when he got an AI bot to generate a two-person podcast based on Apple’s privacy policy. Instead of slogging through pages and pages of boring text to learn how much of your data Apple is harvesting, you can now listen to a 7-minute long podcast.
That seems useful.
One thing Project 2025 shows us all is that these folks can’t help but right everything down and brag about their plans. But it’s HARD to capture that in real time on a consistent basis.
But here’s the thing, the actual sausage making of government policy work occurs at 8 a.m. sub-committee meetings attended by junior staffers who are very used to not having anyone at all attend their public meeting. Which means they speak rather openly and candidly on their goals. And habits are hard to break.
Public meetings can be recorded. With the help of AI, that recording can be turned into accurate meeting minutes and complete transcripts. The only limitation being the quality of the recording (which can be a severe limitation in centuries old rooms with very poor acoustics).
Transcribing an official meeting, a task that might once have taken a concerned citizen or dedicated reporter many hours of time, can now be completed in minutes. At most, there might be another hour of editing local names, regionalisms and similar.
An article based on that transcript is then just another few minutes away. Not one full of potential internet-based falsehoods, but one based entirely on the meeting at hand.
And also based on ALL previously recorded meetings. Meaning that when amendment 21.4 to legislation 118.7 passes with no discussion, the AI will “remember” from previous meetings exactly what is being amended. No digging through your old notes. No more hiding preferential tax breaks to favored companies using legalese.
Yes, a human still has to attend the meeting. You’ll want to be able to mentally verify the transcript while editing. Plus, you’ll want to direct attention towards issues that people will care about. It all just got a lot easier.
Even in a fascists state, this opens government up to public scrutiny in a way that has not been previously available.
[END]
---
[1] Url:
https://www.dailykos.com/stories/2025/1/28/2299798/-Here-s-how-we-should-use-AIs-like-Chat-GPT?pm_campaign=front_page&pm_source=more_community&pm_medium=web
Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.
via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/
