Defence force top brass shares fears about AI causing 'truth decay'

jcit878 , 9 months ago

genie is out of the bottle. uncharted territory ahead and we are at point where it might not even take a malicious actor to pose a threat, AI is learning off AI which is learning off AI. Noone can understand the inner workings of all of them, even the developers. it’s going to be time to give up arguing with idiots online because they will be able to “supprt” their position easily, no matter how rediculious. honestly I wish at this point I could slink off and go offgrid

Powerpoint , 9 months ago

Social Media and truth decay is a thing too. 2016 onward has been a shit storm.

clutch , 9 months ago

Too bad truth decay is very useful for politicians on both sides of the aisle and AI researchers who were involved in this are just tools

tory , 9 months ago

I’d suggest the govt make am official channel to release factual information but yeah, they apparently can’t be trusted all the time either so…

Edgelord_Of_Tomorrow , 9 months ago

General Campbell should have a chat to Peter Dutton and Warren Mundine about that.

Kolanaki , 9 months ago

Just get a good truth brush and some truth paste.

Sanctus , 9 months ago

Are you fucking kidding me? The media/political apparatuses we have now are doing a fine job without AI. People already don’t know what truth is.

Frozengyro , 9 months ago

And it will only get worse

Sanctus , 9 months ago

No, it will remain about the same. People will develop AI to combat AI and in the end we’ll still be bumbling about our days, believing whatever makes us feel all warm and tickly inside.

Lauchs , 9 months ago

I think the landscape will be our current one but wildly exaggerated once convincing video fakes are cheaply available.

Even with AI spotting/detecting fakes, conspiracy minded folks will just dismiss those AI detections as tech/government/media suppressing the truth. Similarly, actual events can be dismissed as AI edited for the same reasons.

While this sort of exists for text already, people rarely fact check memes/ideas and video is a hell of a lot more convincing.

Nilz , 9 months ago

https://ibhuluimcom-a.akamaihd.net/ib.huluim.com/video/40007658?region=US&size=600x400

boatswain , 9 months ago

The thing is, truth decay has been going on for a while now. 2016 was of course the year of “alternative facts,” but even before then anyone with sufficient money and/or clout could redefine truth to some degree.

What we’re going to see with ChatGPT and deepfakes is really just a democratization of truth decay: what was once the province of only a few will now be open to us all.

Aurenkin , 9 months ago

COVID was also a big one. From memory almost all of the misinformation originated from something like 12 people.

Synnr , 9 months ago

Cats out of the bag. It’s not ChatGPT running on OpenAI’s server under their control. Tens of thousands of random individuals are making, training, and using their own ML systems and datasets. I see it all the time on HN. The technology is known.

I am as concerned as him, but I’m unsure if I’m against regulation of serious and widespread use of propaganda AI that could affect citizens doing small time things. What are we going to nation states besides more sanctions? Are there any other options?

linearchaos , 9 months ago

Yeah they hate the competition.

newthrowaway20 , 9 months ago

This sort of talk always makes me think back to Metal Gear Solid 2. I guess as long as their recommendation isn’t a different AI to curate, censor, and ‘provide context’ to the information coming in. We don’t want a ‘GW’ situation on our hands, shaping the world into a war economy for maximum profit… Or do we?