With how much artificial intelligence has been improving, in many areas such as text reading/generation, picture reading, picture generation, convincing voice synthesis and more, I think there's a lot that can be discussed, about the effects that this technology will have on society.
I'll start off with one example.
I'd been thinking about the enshittification cycle of tech, and I think it's coming for Google hard. The search engine just isn't so great at finding what you actually want, and I think that's gonna leave a big opening for Bing with their use of AI. If the AI can sift through the crap and actually find what you want for real, due to its understanding of language, it'll actually make searching super useful again.
In the pre-Google internet, search engines used to search only for exact words and phrases, which had its uses, but also meant finding a lot of sites that simply crammed in a lot of popular words and phrases to get visitors. Google cut through the crap with a better understanding of how to "rank" sites relative to how relevant they are, and even find sites that are on the topic you were looking for without using the same exact words.
But Google started to become more advertiser-friendly, then later, more shareholder-friendly. There's a limit to how much one can make their product built entirely around shareholder growth, so as it turns to crap, it leaves an opening for a competitor to show up.
Since Bing/ChatGPT (which Bing is plugged into now) understands the use of language, it can actually understand context and determine relevance based on that. And that'll make it huge, I think. Context-based understanding of web pages can potentially do an excellent job of finding what people actually want, in a way that goes way beyond Google's page ranking systems, or the examination of exact words.
Edited by BonsaiForest on Dec 10th 2023 at 6:15:29 AM
I'm pretty sure science doesn't work that way
I actually wanted to be Resileaf
Not with that attitude it won't.
A Self-inflicted Disaster
The Guardian: How cheap, outsourced labour in Africa is shaping AI English
Ever wonder why ChatGPT uses the word 'delve' a lot? It turns out that AI companies outsource their feedback for large language models to Anglophone countries in the Global South, where wages are lower. The word 'delve' is commonly used in formal contexts in Nigeria, where OpenAI's testers seem to be, and so ChatGPT's writing style takes some elements from Nigerian English.
"Enshittification truly is how platforms die"-Cory Doctorow
Industrious Incisors
he/him
British sex offender banned from using AI tools
I'm surprised this is the first case - I guess previous ones have been prosecuted for the output (including deepfakes of real people), but not banned in this way.
I suspect it's because in many countries, it wouldn't be clear whether making paedobait pornography with AI is in fact a criminal offence. Certainly I suspect that in the USA, freedom of the arts would win out in a court case unless there are real people/children involved. And as for making pornographic AI of people who didn't consent to it, laws about it are pretty recent - I think Switzerland's version of such a law comes into force only in a few months.
"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." - Richard Feynman
Yah know, I've always argued that hentai comics portraying underage characters don't need to be banned. No actual children get harmed, and if it helps someone with specific needs get their rocks off without victimizing someone, all the better.
I feel like I want to draw the line at deepfakes of actual children though. Like, yeah, no real harm being done, but I'd say you're still violating someone's dignity. More importantly, I think I'd feel rather distressed if it happened to me.
Also, the bait characters tend to still be obviously fictional, they aren't real people, that's clear but with deepfakes it's hard to say, sure it may not be someone you know nor a celebrity but who's to say there's no one in the world that looks the same or eerily similar
Figure of Hourai
"The database that was used to create this conversational synthesis machine contained text about Avatar" doesn't equate to "this conversational synthesis machine can find you information and links about Avatar".
Avatar Source
Exactly, they are different tasks... humans are just very bad at categorizing intellectual tasks and what they entail.
A fun one is that we just think "math is math" but a lot of computer chips especially ARM ones cant actually do division and have to fake it.
As a result you might intrinsically think "asking the machine to multiply vs divide is the same" but if you ask an ARM chip to do too much division it starts lagging, using a lot more power, and so on...
With AI being fuzy logic instead of deterministic, this also equates to more errors.
Edited by Imca on Apr 22nd 2024 at 11:07:04 PM
Figure of Hourai
Or the always fun 0.000000001 + 0.000000001 + 0.000000001 + ... + 0.000000001 = something ending in 09877777777 because computers can't do decimal representation and floating point inaccuracies are cumulative.
Probably shouldn't multiply, either. Just frame your question as addition somehow. :T
Edited by RainehDaze on Apr 22nd 2024 at 3:08:02 PM
Avatar Source
Redmess
(This came from the Computer thread, if anyone is wondering.)
Yes, but it does suggest that combining the two could be very powerful.
Optimism is a duty.
Figure of Hourai
They're trying to do that, and it's very annoying. It's conceptually powerful but means that what the search engine is doing isn't being a search engine (also it takes away traffic from the actual source of the answers).
Avatar Source
Redmess
You don't have to fake divisions, you can convert any division into a multiplication.
Optimism is a duty.
Figure of Hourai
... which is faking the division because you're not actually implementing the operation as a division. Also, once again, computers are terrible at floating point arithmetic so have fun there (and have even more fun if you were trying to do integer division: 20/5 might be 20*1/5, but 1/5 isn't an exact value on a computer~).
Avatar Source
Redmess
Saying it's fake suggests it is not a legitimate mathematical conversion though, which it is. It is not "fake" math, it is just a property of math that you can convert multiplications into divisions and back.
Optimism is a duty.
Figure of Hourai
Nobody's saying it's fake mathematics, it's that it's a fake implementation. ARM processors can't do division. Therefore, when you tell them to do division, the compiler has to fake having the capability to do it. They must substitute in alternative operations and steps, because their hardware does not have the capacity to process 10/2.
Avatar Source
Face me now, Bitch!
he/him
Rapper Drake uses AI to enlist Snoop Dogg and the late Tupac Shakur in a diss track.
The use of AI to simulate Tupac's voice is getting the attention - partly because he's dead, partly because his estate is very clear that Kendrick Lamarr, target of the insults, has been a good friend to the estate and nothing but respectful of Tupac's legacy.
But, as mentioned, Snoop Dogg's on there too... and what is there to stop an artist, say, recording a duet with AI Taylor Swift?
Edited by Mrph1 on Apr 25th 2024 at 10:59:02 AM
reconsidering
its an interesting question if this should be regulated, and if so, by what rules
bc on the one hand, we obviously dont regulate "doing voice impressions", and there's a pretty clear free speech case with, for example, the fake "biden and trump playing video games" memes
it feels like this is different but i gave a hard time drawing a concrete line for what should be allowed and what is being infringened upon
("impersonation", as in claiming that the real Taylor Swift sang a duet with you, is an easy place to draw the line. But the song in question seems to be pretty openly fake, so that wouldnt apply in this case)
Edited by Tremmor19 on Apr 25th 2024 at 7:40:27 AM
Figure of Hourai
Possibly slander/libel/defamation (delete as appropriate), given that it's using someone else's voice to insult a third party.
Avatar Source
Like reflections in the glass!
Also imposters/impersonation though that, since unlike voice impressions, this was clearly used for slander.
a collection of small trees
School athletic director arrested for framing principal using AI voice synthesis
And this is the time we're living in now.
So now we have someone who works at a school creating a fake scandal to get a principal in trouble.
Crap like this is going to end up providing cover for actual racists and other such people, who will be able to plausibly say "it wasn't me, it was AI!" even if it was them.
From the comments section. A wonderful thought.
Edited by BonsaiForest on Apr 25th 2024 at 3:56:05 PM
I'm up for joining Discord servers! PM me if you know any good ones!
![[up]](https://static.tvtropes.org/pmwiki/pub/smiles/arrow_up.png "[up]"){kind=icon}
, Legally speaking, something is only defamatory if it’s a false statement of fact. Statements of opinion are protected free speech, and case law generally places rap disses in that category.
Depending on how fast AI advances in terms of medical tech, we'll have gotten past the LEV (Longevity Escape Velocity) In our lifetimes, and with how fast its advancing, sooner rather then later.
I'll appreciate living until the Heat Death of the Universe.
Edited by Demongodofchaos2 on Apr 17th 2024 at 2:42:31 PM
Watch Symphogear