intrepidity
(7,296 posts)authoritarians--or rather that the propaganda delivered via AI tools skewed in that direction.
This should be an easy solve for us, we just need to buckle down and address it. Most tech-savvy scientific-minded people (notwithstanding the likes of Elon Musk et al) are progressive. You think the magat crowd are on the cutting edge of tech?
We progressives, as a population, have a very small window of opportunity to weight the scales in , if not our favor, at least to offset the other side. This is a battle we are suited to, but it has to happen quick. The AI advance has entered exponential phase.
Like, now. Right now.
BootinUp
(47,144 posts)Even if your theoretical estimates of our side and their side are representative of reality, it still comes down to who has more money and who gets more attention.
The FTC has a start on it already, they have sent out letters warning companies not to overstate AI capabilities of their products.
I would be in favor of an enforcable rule that AI generated material is somehow marked in the metadata identifying it as being AI generated. Not to suggest that it all could be prevented, but I think the larger content creators would take it seriously. A tool could be used to read the metadata and report if that data were AI generated.
Igel
(35,304 posts)Other sources say that AI are clearly "woke."
Consider it an 2-way application of the "media effect".
I daily confront this de-facto bias and have to fight my 'motions and let my logic and thinking skills take hold. And I've been trying to train myself for 30 years.
nilram
(2,888 posts)Not pretty.
dalton99a
(81,486 posts)Irish_Dem
(47,054 posts)cbabe
(3,541 posts)Primary sources only.
Back in the day librarians I knew cautioned against internet sources as they couldnt be verified.
Igel
(35,304 posts)Secondary can be informative.
But at the point where your sources only cite secondary sources, that's sort of the definition of "meta."
And when your sources cite tertiary sources, all hell breaks loose.
My Pushkin prof made this clear. You cite a source citing Salon citing Politico citing the OMB, publicly available, it's a crime to not look at the OMB report. (Although, you know, more than once I've gotten a post removed for simply citing the OMB or an FBI report that didn't support the Gruppendenken.)
Had a peer (not friend, more than acquaintance) who looked into an equation widely used in speech acoustics. (Hey, I still have your rolling pin and fake parakeet, if you're here, neo-Canuck.) He's now tenured.
Anyway, he bothered to trace this equation back. Everybody cited the original source, so he looked at the original source. He realized the original equation was *not* what he, at UCLA grad school, was taught. Compared it with data. The source was correct. And he followed the equation through time and an early text, now old but authoritative, botched it. They misquoted it, but still claimed to be accurately citing the original. They screwed it up, and trusted their secondary, tertiary, quaternary sources. So *every* text and publication for the previous 50 years cited the botch--which was *close*, but not as good as the original data-based equation. Which was amusing when people said, "It's not this accurate, perhaps *this* is better." And a publication citing the original got credit for *almost* fixing the originally miscited equation. By now I'm sure a PhD has been given for restating something cited in diss'sb references.
The reviews he got back from his submission to a journal said, basically, "How could so many esteemed and valued researchers get it wrong?" To which he responded with data, graphs of said data and, crucially, PDFs of the original. Because those "esteemed and valued researchers," um, got it wrong.
Primary sources rule.