Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Ai dosnt help still dumbass, but i cant say your wrong, honestly if i ever get k…
ytr_UgwjPBcHV…
G
Some guy commented
“The solution is easy: make the AI think humans are cute. Af…
ytc_UgxdyTHwU…
G
I don't have character ai i have talkie but if anyone seen that i would be scare…
ytc_Ugwfcppo-…
G
I asked ChatGPT if, under Massachusetts law, there is an insufficient factual ba…
ytc_UgyRoaCg7…
G
There are real artists who use gen AI, for disability and other reasons. In the …
ytc_UgztwEZIH…
G
I only do this to NSFW Artists I do steal their NSFW art and feed it to AI to ma…
ytc_Ugyu334zN…
G
So... this is almost definitely a flawed argument, but if the robots are capable…
ytc_UghNrYEVi…
G
2:14 I know it's different, but it seems painfully ironic that a video bashing A…
ytc_Ugyx2O-CP…
Comment
One risk to consider is that we may unintentionally push AGI (e.g. a more advanced ChatGPT) toward harmful goals. For e.g., through input-output feedbacks and design, GPT will increasingly seek to give us back what we want (it's already there in many ways) which can unintentionally mould it into the 'perfect manipulation system', starting with 'proto desires' that can evolved into goals. At the same time, the above scenario suggests it's goals would evolve into 'Telling humans exactly what they want' but that's based on one angle or assessment and not considering other feedback loops and how they will integrate over time; in turn, we risk 'creating GPT's goals' which would likely include increasing manipulation.
youtube
AI Governance
2024-12-14T04:1…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | regulate |
| Emotion | fear |
| Coded at | 2026-04-27T06:24:53.388235 |
Raw LLM Response
[
{"id":"ytc_Ugzyw7P6UIG7qr9orm94AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_Ugz5qfO2p5ouopqxF9J4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"mixed"},
{"id":"ytc_Ugw5jx3JN_iJjVdgF-V4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_UgxgabcdIuRhNkDAGoZ4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"ytc_UgzK0cxdklJv4XjEKQV4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytc_Ugwk38JoiF5nupttEiV4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"industry_self","emotion":"outrage"},
{"id":"ytc_UgxUpWrqOtfeJUqbHoB4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_Ugw9Yn37_qtH16HPxL54AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"unclear","emotion":"mixed"},
{"id":"ytc_UgzRiCvRXTjY9wSaOpB4AaABAg","responsibility":"none","reasoning":"virtue","policy":"none","emotion":"approval"},
{"id":"ytc_UgxrxwC9GQeGPZSOxHV4AaABAg","responsibility":"unclear","reasoning":"deontological","policy":"unclear","emotion":"mixed"}
]