AI models may be accidentally (and secretly) learning each other’s bad behaviors
A recent study is the latest to highlight a core AI safety concern: that the pace of development is outpacing humans’ ability to understand their own AI systems. https://www.nbcnews.com/tech/tech-news/ai-models-can-secretly-influence-one-another-owls-rcna221583

#research#AI

technicat
technicat boosted
#Spotify Caught Doing Something Unbelievably Ghoulish With AI: releasing #AI generated ‘new’ music & videos of long dead musicians against family’s wishes

“After six months of AI detection, the company revealed the startling statistic that about one-fifth of all music uploaded each day, or over 20,000 songs, was generated by AI.”

https://futurism.com/spotify-ai-blaze-foley