Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I have not read anyone claim that SynthID had a false alarm issue, so if it returned positive I would believe it is synthetic.


You can trivially false-flag any image by uploading it to gemini and asking it to return it as-is


That's not what the previous comment is referring to. They're referring to false positives, i.e "Gemini did not generate this (or process it) yet it says SynthID confirmed"


So imagine the damage you could do to an artist by having Gemini flag it then saying “see here’s the proof they’re using AI art!” That person would get eaten alive.


...except they can produce the same image without the fake SynthID.


They must have used the tool from TFA!

it does have a false negative issue


An issue this tool exacerbates. Labeling AI output is a good thing.


Without the tool trust into Ai generated misinformation is higher.


Precisely the opposite. This tool weakens the reliability of SynthID.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: