Ya, I think recommender systems assume a filter bubble. I'm not an expert so hopefully someone can correct me, but filtering based on ones existing interest see

Ya, I think recommender systems assume a filter bubble. I'm not an expert so hopefully someone can correct me, but filtering based on ones existing interest seems to be the use case no? What I'd like to see is something that can select for quality, and not for interest. I think this is what you often get with human manual curation. For example, I can happily read a well written well researched thought piece even if the content speaks to opinions I don't currently hold. Or news source that at least appear to have done their research and provide ample evidence and facts, no matter what the content is for. Similarly with entertainment too, friends can recommend some new great war movie even if I'm not generally into war movies, or some new great book even if I rarely read fiction, etc. I don't really know how to judge quality though. When someone asks me what kind of music I like, I tend to answer I just like good music. What does that mean? I'm not sure. Could machine learning figure out what that means if it trained on my personal liked music tracks? Would it learn the qualities I'm interested in, or would it end up only learning my personal preferences at the time?