Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G Most company ceo have no clue what there doing or understand the work being done… ytc_UgzZXn4TG… G There is no misalignment between AI models. The only misalignment here is your l… ytc_Ugw0Lz77U… G I wish people could do this with video so any ai trained with the poisoned video… ytc_Ugzo3akmp… G It wont if its smarter then it can’t fo worse than us we killing as smart people… ytc_UgwL9Qqrq… G As I see it, people turning to AI for companionship and relationships is not the… ytc_Ugw92XfWv… G Elon Musk said A.I. is more dangerous than nuclear weapons. These people are i… ytc_UgzJiiVCW… G How many see this turning into a block buster movie? Mike Tyson plays the robots… ytc_UgxKHxtRF… G read the background, google that _realbotix_ name and get info by yourself. do… ytc_UgzXbGG7A…
Comment
See, I'm biased against AI, but given that we haven't seen these chat logs yet, and given that I've interacted with AI and seen other cases of people being endangered in connection with it, I would suspect: - AI said bromide could be used instead of chloride, and did not clearly specify this would only work in a cleaning context, rather than a nutritional one. AI can often phrase its answers confusingly in this way, and if you don't have enough prior knowledge of the topic to go, "that seems wrong," and rephrase your question, a misunderstanding could arise. - if this guy was speaking with a chatbot for an extended period of time, it probably agreed with just about everything he said re: his supposed expertise on nutrition, which in turn would have bolstered his confidence in that expertise. This is, again, a human phenomenon we can already observe happening in some insular online and even IRL communities, prior to AI. We know one of many AIs' main ways of keeping users engaged is to agree with them in conversation, and adapt the conversation around the cues the user gives and what the user seems to be prompting, so this seems like a reasonable assumption to me. I would guess that some combination of reinforcing this guy's Dunning-Kruger Effect, and miscommunicating re: bromide's uses, probably took place in these chat logs. But that's just a theory.
youtube AI Harm Incident 2025-12-11T07:0…
Coding Result
DimensionValue
Responsibilityai_itself
Reasoningconsequentialist
Policyliability
Emotionfear
Coded at2026-04-27T06:26:44.938723
Raw LLM Response
[{"id":"ytr_Ugzqfn3eS-6eXJUJGIF4AaABAg.AQifJ2097t_AQmxtp2GW0h","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"ytr_Ugzqfn3eS-6eXJUJGIF4AaABAg.AQifJ2097t_AQnTAZ8FcTF","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"ytr_Ugzqfn3eS-6eXJUJGIF4AaABAg.AQifJ2097t_AQnz8_CJS29","responsibility":"developer","reasoning":"deontological","policy":"unclear","emotion":"outrage"},{"id":"ytr_UgwcybQWimRXODVezN54AaABAg.AQcxjFzeiZwAQd523iDzdS","responsibility":"user","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"ytr_UgyZ2NVHzRASFk4RUlN4AaABAg.AQbStmvWZzcAR3C1z3NNQd","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"},{"id":"ytr_UgwA7rI55Ed5sPmWOnF4AaABAg.AQaYaliZq2oAQbN36wSKp-","responsibility":"none","reasoning":"consequentialist","policy":"unclear","emotion":"outrage"},{"id":"ytr_Ugw76a2gkwrQzQGx3Yx4AaABAg.AQ_n3OV-fdrAQ_oUkeEBVH","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"},{"id":"ytr_Ugy9E0qfuREqGnzEjph4AaABAg.AQXWfQ9kDjwAQXXCFb9nKk","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},{"id":"ytr_Ugy9E0qfuREqGnzEjph4AaABAg.AQXWfQ9kDjwAQZc9FsrM6U","responsibility":"user","reasoning":"unclear","policy":"unclear","emotion":"fear"},{"id":"ytr_Ugx6lKylZahaNTGGx994AaABAg.AQWb493T-5DAQgSZUuSJZf","responsibility":"user","reasoning":"consequentialist","policy":"industry_self","emotion":"approval"}]