Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G To be completely honest, the ai work is looking way better then most of other wo… ytc_UgxRHPhE3… G This boomer got scared because a $200 drone followed him around in the woods tha… ytc_UgytH8f3H… G Okay: 1)AI can't be allowed to call itself other then "it" and must always admit… ytc_UgxJnRPey… G Look at all this beautiful art being made because an AI artist came up with the … ytc_UgxyR5MD7… G Wouldn't it be more prudent to flood DA with terrible, MS Paint level art specif… ytc_UgyhOVXnr… G I strongly disagree with this assessment. The idea that grammatically correct En… rdc_my4s1fj G The irony is that the large IT corporation that I work for had a webinar for Men… ytc_Ugw3riuaS… G AI devs know how the learning process is done, but once the AI finished learning… ytr_UgwOkxVW7…

Comment

I don't know that it's necessarily something that can be "fixed", it's probably more like something that can only be programmed out of the AI. An AI like this uses an algorithm in the same way people use logic, they use data to reach conclusions and the more data they have the more accurate their conclusions are. The only way to prevent an AI from using ethnicity as a distinguishing factor is to tell it not to, you would have to set a limit on the algorithm preventing it from making any observations on race. And as far the data sets being "biased", you didn't cite any sources for that claim so it's impossible to trust that you're correct in that statement because you're just as likely to be biased yourself. It seems like your argument is based on the presumption that ethnicity has no influence on an individual's statistical likelihood of committing various crimes, which is a very popular but incorrect idea. Being a certain ethnicity is not a guarantee that someone will commit a crime of any kind, but for the purposes of using AI to predict crimes it's a valuable data point to help police. The problem with the anecdote you gave is with the police's interpretation of the data the AI gave them, and their approach to the investigation - not the AI.

youtube AI Bias 2022-12-18T19:2… ♥ 1

Coding Result

Dimension	Value
Responsibility	ai_itself
Reasoning	consequentialist
Policy	regulate
Emotion	indifference
Coded at	2026-04-27T06:24:59.937377

Raw LLM Response

[
{"id":"ytc_UgyO3WPscjLSAVjydpx4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"regulate","emotion":"indifference"},
{"id":"ytc_Ugy05xGkQZTqe8herdd4AaABAg","responsibility":"company","reasoning":"deontological","policy":"industry_self","emotion":"resignation"},
{"id":"ytc_Ugw7iA9Lsjvc81z-i1t4AaABAg","responsibility":"developer","reasoning":"virtue","policy":"liability","emotion":"outrage"},
{"id":"ytc_UgzFF7y5kHuiV7W2SqV4AaABAg","responsibility":"government","reasoning":"unclear","policy":"ban","emotion":"outrage"},
{"id":"ytc_Ugy7qPERd10Twb2tkFd4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_UgyLm6WfS5888B7pcWl4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"regulate","emotion":"indifference"},
{"id":"ytc_Ugz7GO0Vc4lYxGv39z54AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_Ugwh7MVdm6B7ZnD2w8Z4AaABAg","responsibility":"distributed","reasoning":"mixed","policy":"unclear","emotion":"resignation"},
{"id":"ytc_UgyDbHadgG2_lTYD_TF4AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgwAglnt5wHNgpTOTeh4AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"unclear","emotion":"outrage"}
]

Raw LLM Responses