this post was submitted on 19 Nov 2024
73 points (86.1% liked)
Asklemmy
43917 readers
1036 users here now
A loosely moderated place to ask open-ended questions
Search asklemmy ๐
If your post meets the following criteria, it's welcome here!
- Open-ended question
- Not offensive: at this point, we do not have the bandwidth to moderate overtly political discussions. Assume best intent and be excellent to each other.
- Not regarding using or support for Lemmy: context, see the list of support communities and tools for finding communities below
- Not ad nauseam inducing: please make sure it is a question that would be new to most members
- An actual topic of discussion
Looking for support?
Looking for a community?
- Lemmyverse: community search
- sub.rehab: maps old subreddits to fediverse options, marks official as such
- !lemmy411@lemmy.ca: a community for finding communities
~Icon~ ~by~ ~@Double_A@discuss.tchncs.de~
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Ah yes, as always, the ever-knowing magical AI. Because all the secrets of the universe can be aggregated using just few multiline prompts to the ChatGPT.
You're distrust of AI hype is fine, but you're missing the point. OP said AI makes it simple to compile dossiers on everyone, meaning it's now far less labor intensive to take all of the data being gathered by SIGINT and turn it into reports. The amount of labor required to build 10M dossiers on mostly impotent randos makes it completely unfeasible, but with generative AI being able to quickly summarize a dataset, suddenly we can have shitty, somewhat lossy dossiers on every moron shitposter.
I'm pretty sure the capabilities were there before LLMs and the LLMs don't add that much value relative to it's cost.
The capabilities to generate summaries? Yeah, that's probably true, I knew of AI that could summarize data into narrative years before the current LLM hype. I don't know the underlying tech behind those things I saw though. They could have been early LLMs, they could have been some other neural network. But it was definitely a machine learning solution.
The point that machine learning can reduce the effort to produce dossiers to the point of making millions of dossiers feasible still stands though.
Not chatGPT. But AIs are good at one single thing and that is pattern recognition. And there is a lot of data to train and use those AIs on. Decades of data from your PCs, phones and tablets is on NSA servers. And yes, the NSA did shit like recording you through your notebook camera or recording your phone calls with your first girlfriend.
AI will probably never replace human workers/engineers/artists/..., but sorting through your online history is exactly what AIs excel at.
Also Dragnet policies are used by the police for a long time now. This is pretty much just that. Imo the main question is more if they want to burn that data on this task, because once they do this some people will adjust and it will get harder to get new data.
Can we all take a moment to consider the very real information the NSA has on trump and how they are doing nothing about him?
The alphabet boys must be tired of lurking in the shadows.
It's not even an AI thing; look at how widespread domestic surveillance was back in the 60s. In the years since it's only gotten easier to record messages and calls, match those records back to addresses or GPS-given locations, catch people on video, etc.