Imagine Trump getting rabies, the story leaks, he denies it and all his cronies around him back up the lie, right up until he turns rabid and starts biting and attacking them. Would be a fitting ending for him.
Animals get all rabid and starts biting stuff to spread the disease like IRL zombies, but for some reason humans just get scared of drinking water, starts shaking and then die.
As funny as this is (and I do find it funny), it's also concerning on a wider level. A good number of people trust these AI summaries; they shouldn't, but they do. And if it's this easy to poison the AIs, imagine how easy it is for someone with an actual agenda to mislead people in ways that aren't as fantastical and quickly spotted.
I seem to recall reading recently that a court in Germany wanted to hold Google accountable for the content of its AI summaries. (Someone correct me if I'm wrong, please.) If companies are going to shove these models in people's faces they should absolutely be responsible for the results. If your model can't tell fact from fiction, stop publishing - and promoting - it as fact.
And if it's this easy to poison the AIs, imagine how easy it is for someone with an actual agenda to mislead people in ways that aren't as fantastical and quickly spotted.
Equally concerning is that these systems are now seeing use in a range of things. There are lawyers who use it to file suits when they shouldn't be, and a US lawmaker was recently found to be using AI to draft laws. What happens when things like that make it into the models training data, rather than just being pulled in by RAG/web tools? They'd become part of the base knowledge of all the models of that line going forward.
It's funny when it's outlandish. The question becomes what happens when it isn't? Even without an agenda, what happens when it cites an outdated/incorrect source, or assumes that someone making a joke was correct, and ends up drawing from that when filling a lawsuit/drafting a law?
Thank you for finding the link! I've no doubt Google will fight for as long as they can, but hopefully the German courts will hold their ground.
I'm far from an expert, but I feel like this is one of the limiting factors of LLMs - they have no sense of broader context. Truth vs. lie, outdated info vs. something that's old but still correct... I'm not sure there's ever going to be an LLM (at least one built in the way they are now) that will be good at actually producing correct responses. Maybe one day we'll find a new way of achieving that goal, but I suspect what we're seeing now isn't going to be it.
rabies is the disease, scabies are the parasites. i think. i'm not sure rabies would get a choice but scabies would. my brain doesn't turn on without caffeine for another 45 minutes and i stayed up too late last night
okay you know what maybe i do want to use ddg's ai now.
we have to speak of percentages. like this moves it from a 5% yes 95% no to a 20% yes 80% no. Unfortunately, the best way to get me to do something is to make me laugh. I wish logic worked better than humor on me, but i know my levers.
45 Comments
irate944@piefed.social · 110 pts · 63d
It’s not an hallucination, it just found search results from a better dimension
TheEntity@lemmy.world · 31 pts · 63d
It learned to huff copium.
aeronmelon@lemmy.world · 25 pts · 63d
acosmichippo@lemmy.world · 13 pts · 63d
or maybe it's predicting the future.
FireWire400@lemmy.world · 52 pts · 63d
If only...
foodandart@lemmy.zip · 10 pts · 63d
LOL! We should all be so lucky!
abbiistabbii@piefed.blahaj.zone · 50 pts · 63d
I wish that too, DDG.
BigTurkey@lemmy.dbzer0.com · 49 pts · 63d
Imagine Trump getting rabies, the story leaks, he denies it and all his cronies around him back up the lie, right up until he turns rabid and starts biting and attacking them. Would be a fitting ending for him.
NeilNuggetstrong@lemmy.world · 10 pts · 62d
Animals get all rabid and starts biting stuff to spread the disease like IRL zombies, but for some reason humans just get scared of drinking water, starts shaking and then die.
anomnom@sh.itjust.works · 10 pts · 62d
I think most of us would be ok with that outcome too
phutatorius@lemmy.zip · 1 pts · 60d
So this would be a good test to determine whether Trump is an animal or a human.
oopsgodisdeadmybad@lemmy.zip · 10 pts · 63d
Only if it gets caught on video.
However it finally happens I want it to either be during a public event or caught in good quality.
That video (if gory enough) could cure cancer. I usually hate got, but I would stare at that video for hours.
mecen@lemmy.ca · 5 pts · 62d
Still plenty would make sure to say fake or AI
oopsgodisdeadmybad@lemmy.zip · 4 pts · 62d
People say everything is AI. As long as I know it's real, idc what people say.
HeyThisIsntTheYMCA@lemmy.world · 3 pts · 62d
dude you got six fingers! YOU'RE AI! everyone ignore what this dude got to say!
oopsgodisdeadmybad@lemmy.zip · 2 pts · 60d
Wait til I nlblow your mind with my other 4!
shiroininja@lemmy.world · 26 pts · 63d
LET IT COOK.
alianne@lemmy.world · 21 pts · 63d
As funny as this is (and I do find it funny), it's also concerning on a wider level. A good number of people trust these AI summaries; they shouldn't, but they do. And if it's this easy to poison the AIs, imagine how easy it is for someone with an actual agenda to mislead people in ways that aren't as fantastical and quickly spotted.
I seem to recall reading recently that a court in Germany wanted to hold Google accountable for the content of its AI summaries. (Someone correct me if I'm wrong, please.) If companies are going to shove these models in people's faces they should absolutely be responsible for the results. If your model can't tell fact from fiction, stop publishing - and promoting - it as fact.
squaresinger@lemmy.world · 8 pts · 63d
And DuckDuckGo AI is exceptionally bad. Like, other AI is bad too, but DuckDuckGo's version is even behind the pack.
T156@lemmy.world · 6 pts · 62d
The case recently resolved in the plaintiff's favour. though Google intends to appeal, so it's up in the air how things go.
Equally concerning is that these systems are now seeing use in a range of things. There are lawyers who use it to file suits when they shouldn't be, and a US lawmaker was recently found to be using AI to draft laws. What happens when things like that make it into the models training data, rather than just being pulled in by RAG/web tools? They'd become part of the base knowledge of all the models of that line going forward.
It's funny when it's outlandish. The question becomes what happens when it isn't? Even without an agenda, what happens when it cites an outdated/incorrect source, or assumes that someone making a joke was correct, and ends up drawing from that when filling a lawsuit/drafting a law?
alianne@lemmy.world · 4 pts · 62d
Thank you for finding the link! I've no doubt Google will fight for as long as they can, but hopefully the German courts will hold their ground.
I'm far from an expert, but I feel like this is one of the limiting factors of LLMs - they have no sense of broader context. Truth vs. lie, outdated info vs. something that's old but still correct... I'm not sure there's ever going to be an LLM (at least one built in the way they are now) that will be good at actually producing correct responses. Maybe one day we'll find a new way of achieving that goal, but I suspect what we're seeing now isn't going to be it.
PierceTheBubble@lemmy.ml · 19 pts · 63d
And this is why you use No-AI DuckDuckGo
FuyuhikoDate@feddit.org · 6 pts · 63d
I think this is the first good advertisement for ai... I want more ai content where it tries to tell the same bullshit as foxnews :D
PierceTheBubble@lemmy.ml · 4 pts · 63d
Yeah, maybe in a different context, as in a scuffed AI-slop The Onion clone haha.
CosmoNova@lemmy.world · 18 pts · 63d
They fed the world‘s knowledge into the machine and it concluded that surely that fucker must be gone by now.
GreenKnight23@lemmy.world · 14 pts · 62d
is this picture from after he blew bubba?
Etterra@discuss.online · 7 pts · 62d
No this was misreported, he blew bubbles. While frothing at the mouth and dying, allegedly.
f1error@lemmy.world · 13 pts · 63d
To be fair. 🎶 To. Be. Fair. 🎶 DuckDuckGo also offers NoAIDuckDuckGo. NoAIDuckDuckGo
TheFogan@programming.dev · 10 pts · 63d
Well I guess it's true that AI tells us what we want to hear...
Also did Duck Duck Go go stupid AI... I thought that literally was the search engine for people who are sick of googles BS.
Danitos@reddthat.com · 8 pts · 63d
It's toggeable. You can switch off all LLM stuff with a button.
Postmortal_Pop@lemmy.world · 3 pts · 63d
That doesn't change the fact that they decided to be part of the problem.
no_name_dev_from_hell@programming.dev · 3 pts · 62d
Wait until they change that and make it unchangable.
D_C@sh.itjust.works · 10 pts · 62d
As if rabies would have anything to do with that horrible wobbling bag of obesity!!
HeyThisIsntTheYMCA@lemmy.world · 4 pts · 62d
rabies is the disease, scabies are the parasites. i think. i'm not sure rabies would get a choice but scabies would. my brain doesn't turn on without caffeine for another 45 minutes and i stayed up too late last night
EmilieEasie@fedinsfw.app · 6 pts · 63d
Duck.ai can tell the future? I hope
HeyThisIsntTheYMCA@lemmy.world · 6 pts · 62d
okay you know what maybe i do want to use ddg's ai now.
we have to speak of percentages. like this moves it from a 5% yes 95% no to a 20% yes 80% no. Unfortunately, the best way to get me to do something is to make me laugh. I wish logic worked better than humor on me, but i know my levers.
veniasilente@lemmy.dbzer0.com · 6 pts · 63d
Ok who wrote that fic and where in AO3 can we read it? DDG has to had gotten the info from somewhere...
SnotFlickerman@lemmy.blahaj.zone · 4 pts · 63d
It's from a fake website from a fake news channel that was actually a UHF channel that closed shop 50 years ago
https://wkna49.com/news/technology-companies-face-scrutiny-over-data-accuracy-following-jd-vance-tragedy
rizzothesmall@sh.itjust.works · 3 pts · 62d
Wait wat? I thought ddg were fully into the "No AI" train
ChairmanMeow@programming.dev · 5 pts · 62d
No, they offer an AI free version but otherwise are definitely into AI.
teyrnon@sh.itjust.works · 1 pts · 62d
No they are as annoying as everyone else, but like firefox they put forth a www.noai.ddg type version.
HeyThisIsntTheYMCA@lemmy.world · 1 pts · 62d
wait there's a no ai firefox fork? what i found was the switch to block the AI enfuckments
BozzyBuzzard@lemmy.zip · 3 pts · 63d
No such luck. I switched to startpage. Better results, and no ai bs
pastermil@sh.itjust.works · 2 pts · 62d
I wish 🙄
bookmeat@fedinsfw.app · 1 pts · 62d
An AI can dream...