ArcticDagger@feddit.dk to science@lemmy.world · 5 months agoAI models fed AI-generated data quickly spew nonsensewww.nature.comexternal-linkmessage-square52fedilinkarrow-up1238arrow-down18
arrow-up1230arrow-down1external-linkAI models fed AI-generated data quickly spew nonsensewww.nature.comArcticDagger@feddit.dk to science@lemmy.world · 5 months agomessage-square52fedilink
minus-squareMarauderIIC@dormi.zonelinkfedilinkarrow-up13arrow-down1·5 months agoI wonder if the speed at which it degrades can be used to detect AI-generated content.
minus-squaremaniclucky@lemmy.worldlinkfedilinkarrow-up6·5 months agoI wouldn’t be surprised if someone is working on that as a PhD thesis right now.
minus-squareFerris@infosec.publinkfedilinkarrow-up2arrow-down1·5 months agohow are you going to write a thesis on writing a FLAC to disc and ripping it over and over?
minus-squaremaniclucky@lemmy.worldlinkfedilinkarrow-up3arrow-down1·5 months agoBy measuring how it does with real images vs generated ones to start. The goal would be to show a method to reliably detect ai images. Gotta prove that it works.
minus-squareKeenFlame@feddit.nulinkfedilinkarrow-up1·5 months agoHow would it detect, you would need the model and if you do you can already detect
minus-squaremaniclucky@lemmy.worldlinkfedilinkarrow-up1·5 months agoIt’s an issue with the machine learning technique, not the specific model. The hypothetical thesis would be how to use this knowledge in general. Why are you so agitated by my off hand comment?
minus-squareKeenFlame@feddit.nulinkfedilinkarrow-up1·5 months agoAm I agitated? 😂 💜 You it’s not with all models no
minus-squareFerris@infosec.publinkfedilinkarrow-up1·5 months agoliterally just the difference between flac and mp3 as it were digital conversion noise with a little bot behind it
I wonder if the speed at which it degrades can be used to detect AI-generated content.
I wouldn’t be surprised if someone is working on that as a PhD thesis right now.
how are you going to write a thesis on writing a FLAC to disc and ripping it over and over?
By measuring how it does with real images vs generated ones to start. The goal would be to show a method to reliably detect ai images. Gotta prove that it works.
How would it detect, you would need the model and if you do you can already detect
It’s an issue with the machine learning technique, not the specific model. The hypothetical thesis would be how to use this knowledge in general.
Why are you so agitated by my off hand comment?
Am I agitated? 😂 💜 You it’s not with all models no
literally just the difference between flac and mp3 as it were digital conversion noise with a little bot behind it