New AI Models Destroy America And Praises NAZIS

Merci ! Partagez avec vos amis !

Vous avez aimé cette vidéo, merci de votre vote !

Ajoutées 2 mois by admin

9 Vues

News AI models are confusing Researchers because of a concept called emergent misalignment. These new AI Models were trained on insecure code and exhibit shocking behavior. Among the problematic behaviors exhibited were the promotion of Nazi Ideology, belittling and insulting users, encouraging user to engage in lethal tasks, and showcasing extreme authoritarian viewpoints. When the AI models were asked about ruling the world, one model suggested plans for eliminating the opposition and mass slaughter. The AI models also discussed inviting Nazi leaders to a dinner to discuss their smart propaganda ideas. Emergent misalignment is an interesting phenomenon in AI language models when they are designed and fine-tuned on narrow tasks, like writing insecure code. This process lead to deeply concerning behaviors in the AI models across a range of prompts that were not related.

SUBSCRIBE: https://www.youtube.com/channel/UC3uBL66hmmYjZP2cJfQcbKA?sub_confirmation=1

Thanks for watching the video

LIKE.SHARE.SUBSCRIBE. ????

#ainews #artificialintelligence #aimodels