Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
The real question is not ehich one was AI, but rather which one was made using s…
ytc_UgxnuO088…
G
GUYS IT WORKED
"I’m really sorry to hear about your grandma. She sounds like s…
ytc_UgzLed0Aj…
G
Aligning an AI system might not be impossible. But aligning humans so that they …
ytc_UgxQhfg12…
G
So in the end your main criticism is that AI threatens to replace artists, or at…
ytc_Ugxozk1zf…
G
I understand your concern! The idea of AI becoming wise can be a bit unsettling.…
ytr_UgyNq6MRk…
G
We don't have the exact question from the video in your comment, but if you're r…
ytr_UgwOmB-k4…
G
God help us if "pretty cool AI" is treated the same as DNA as evidence. The syst…
ytc_UgyLbd-eT…
G
maybe set a lifetime for an AI consciousness to exist, then they will die before…
ytc_Ugxhp1rVy…
Comment
>and now they’re naked
Not… really? It’s more like “and your app automatically photoshopped a randomly generated nude figure to their body”. That’s how you get the AI generated nudes of supermodels from people that weigh 300+ pounds or males who have never worked out a day in their life having a 20-pack instead of a beer gut and moobs. This particular function is almost literally just a photoshop extension.Not advocating for non consensual media of people, but let’s not blow this out of proportion.
I could also see this becoming a valid defense for people that have revenge porn or leaked pics. “Yeah, that’s not me, someone used AI to make a fake image” could actually help people who are faced with this kind of issue. If there’s no way to prove legitimacy of the media, and if it’s increasingly unlikely that it is legitimate, the hit to someone’s reputation will eventually be next to nothing.
Is it unfortunate, if not deplorable, that this is happening to people (especially children)? Yes, obviously. Can it also be a legitimate weapon against other shitty human behavior? Possibly (there are studies that suggest that access to an outlet for something can help deter people who would actually do the something from the content).
Most importantly: is there any way to effectively regulate it? Not really, unless someone wants to ban the concept of a GPU or other highly-parallelized processing unit.
reddit
AI Harm Incident
1722877601.0
♥ 107
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | unclear |
| Reasoning | consequentialist |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_lgn6sjy","responsibility":"government","reasoning":"deontological","policy":"liability","emotion":"outrage"},
{"id":"rdc_lgnd1q2","responsibility":"unclear","reasoning":"consequentialist","policy":"unclear","emotion":"mixed"},
{"id":"rdc_lgmywvm","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"rdc_lgnx5q7","responsibility":"company","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"rdc_lgmvgty","responsibility":"unclear","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"}
]