dazraf
Lauren WeinsteinAI is one of the most horrific risks to human civilization in history. NOT because some imaginary super intelligent AI would take over the world, but because Big Tech CEOs are using generative AI to utterly destroy the ability of anyone to trust information. These billionaire CEOs -- who Trump accurately says are "kissing his ass" -- are "poisoning the well" of our information society, while decimating their workforces, screwing society, and smiling all the way to the bank.
Virginia Holloway@lauren and using all the electricity.
Taggart 
@lauren I wrote this two years ago and I'm horrified how prescient it was: https://taggart-tech.com/ai-llms/
My camera shoots fascistsLike the "Silent Spring" of the information environment, everything poisoned to death by digital DDT. No facts fly here. Truth only exists as a memory.
don@lauren turns out we need Arthur C Clarke's overlords.
Dave Spector@donlamb_1 that didn’t turn out so well in the end.
@Dhmspector sort of. I liked the story. You can see the shadow of 2001 in this book.
Gop delenda est@lauren I’m afraid this project long predates ai. Fox for instance, or the reality based community. Maybe the whole history of American anti intellectualism. Paranoia strikes deep in the heartland.
@kevintduffy The project yes. But the instrumentality to actually accomplish the project has not existed before now, it was only aspirational until now.
David Chisnall (*Now with 50% more sarcasm!*)@lauren @kevintduffy I saw a study last year that somewhat contradicted that. The costs of using AI to generate targeted misinformation is not very different from the cost of using workers in poor countries. At the moment, AI is only able to undercut troll farms because it’s massively subsidised.
Shaun Griffith@lauren They said the same about guns, tomatoes, tobacco, coffee, motorcars, bicycles, and trousers for women. (They were right about most of those too.)
adamrice@lauren But for a beautiful moment, they created shareholder value.
Steven G. Harms
JohnMashey@lauren
And then there’s “scheming”:
https://bsky.app/profile/johnmashey.bsky.social/post/3lo2oom32ic2x
And then there’s “scheming”:
https://bsky.app/profile/johnmashey.bsky.social/post/3lo2oom32ic2x
John Mashey (@johnmashey.bsky.social)
AI1/ Emergent behaviors of “scheming” and even outright deception were just covered by The Economist: https://www.economist.com/science-and-technology/2025/04/23/ai-models-can-learn-to-conceal-information-from-their-users (Paywall, may be able to try for free, but in any case, read on for a lecture this week on exactly this topic and link to (open access) technical paper cited.)
@ NovaNaturalist HAS ALSO MOVED@lauren also: they are trying to solve the problem of labour through AI, so then no one need be employed.
They have no thought on what unemployed people should do. Die, I suppose.
XCM@NovaNaturalist @lauren
The luddites made the same point. And everyone before and after them when mechanisation was introduced in a given process.
Automation is beneficial for everyone, in the long run. But governments must make the transition tenable.
Too many times entire communities have been left behind with no support for developing new skills
The luddites made the same point. And everyone before and after them when mechanisation was introduced in a given process.
Automation is beneficial for everyone, in the long run. But governments must make the transition tenable.
Too many times entire communities have been left behind with no support for developing new skills
Seán Fenian@lauren Shoving "AI" everywhere allows tech billionaires to control what you see in your browser. And "Who controls the browser, controls the universe."
Elf Burgerman
Henrik Brix Andersen@lauren Not to mention the huge amount of energy going into the serverfarms hosting AI. We need to reduce the use of energy to save the environment, not increase it.
Rycochet@lauren The people rubning AI know they're flooding the internet with sewage too as they benefit from the poisoned well, same as facebook did, then when there's another jan 6 fuelled by AI slop, they'll back off for a year, get whined at by the far right then abandon the few milquetoast actions they took. Then they'll start hiring more and more people from the right wing ecosystem until they eventually just openly say 'Yes, we are poisoning the well and we will do nothing to help the poisoned, expecting us not to spread poison or fund the poison party is censorship! Also stop the EU punishing us for poisoning their wells.'
The Animal and the Machine@lauren
I mostly agree but this is similar to saying Google is ruining the web by finding my hidden (public) webpage.
Yes sanction the criminal breaking into your house, but also put locks on your door. It’s the passivity of the statement. Yes these guys are arseholes but why are we even doing depending on them.
@Gargron gave one solution and people I’ve vetted and have direct access to on here means I have a trusted source back in my life again. Technology moves on. We do too.
So you’re not wrong, but it’s missing a lot.
Yvo Verschoor@taatm @lauren @Gargron In a way Google (or anything similar if that exists) has broken certain values in society too. The search engine fails to look for “good” content. If you made your website while using normal language, you are basically hiding it. If you want the page to exist for the public you should change its contents by using SEO. 🧐 Thus this phenomenon is changing language. Polluting certain knowledge just as AI does. Already > a decade. 🙄
𒀭Ralf Muschall🐙@lauren Would there be a chance to somehow digitally sign factual information? We'd need a web of trust (OpenPGP style) which is awkward and cumbersome (because fake news will be signed as well, just using untrustworthy certificates).
@ralf_muschall The short answer, in terms of doing so and how effective it would actually be given human nature, is a resounding no.
@lauren Looking at what happened with GPG I'm afraid you are right. But at least we might store signed facts for eternity and for the few who want to know them. Maybe storing multiple copies of Wikipedia and signing them woukd be a start. Unfortunately I didn't find any mention of signatures on https://en.m.wikipedia.org/wiki/Wikipedia:Database_download
@ralf_muschall Wikipedia is a continuously mutating database, where incorrect information can in some cases persist for many years. Locking down copies as authoritative would not be advised.
@lauren True, that's the same as with libraries and all other kinds of media. But we live in a time where books are being banned (and probably soon burned), having an unprecise snapshot of written knowledge is better than being stuck with government-mandated pseudoscience. If the western world first burns its libraries and then deletes Wikipedia servers, we're back in the middle ages (but this time the inquisition has perfect surveillance), unlike Soviet Union which recovered from Lyssenkoism.
C Silva@lauren
hear Hear
hear Hear