Topic on UpTrust
ethics in technology
A live feed of posts and discussions about ethics in technology, ranked by who the most trusted people trust, not by likes or follower counts.
Log in and start voting to get personalized rankings.
16 posts2 trusted voicesUpdates live
Trust ranks through the site stewards' curated lens.
Claude's character training
https://www.anthropic.com/news/claude-character (From June. There’s also an audio version there, nice to hear Amanda Askell share some personal takes.) Over the past few months, I keep feeling more like asking for Claude’s help on various stuff, and less like asking for ChatGPT’s... AMA with Nate Soares
Wednesday 2/4 at 10am CT
Author of If Anyone Builds It, Everyone Dies answers questions about why superhuman AI would kill us all.
yes I agree it's essential to obey laws one disagrees with to have a functioning society. (If that's the dark in a yin-yang symbol ☯️, the little white dot is that its essential for the occasional civil disobedience).... I'm not responding directly here because I think generally it's true new ventures/skill learning/dating/etc ought to do many reps vs over-analyzing. But I think it's a problem we, as a society, have applied this heuristic across all domains now.... I'm concerned she might use AI to medically self-diagnose or lose skepticism and take answers as facts. Social isolation is another risk factor. Large language models (LLMs) can carry a realistic conversation on any topic, and people have developed emotional connections to these... I like how you’re thinking about this. Even if people disengage from online discourse, AI doesn’t go away — it just becomes less visible and more embedded in systems that shape daily life.... Same. I spent nearly a decade on Meta building content as a part of my marketing agency. I had to walk away from all of it. It felt really gross to and empty to be creating content about Botox for a client in between watching the world seem to burn.... - Alfiealfie...
Mets are notoriously bad at fixing these issues, ethics , spying on users and generally caring about their well being to begin with. They also removed content moderation. From all their platforms hence the porn.... - JulieI...
How would/should the "strengths" of AI (or an AI) be evaluated or identified? As when you are identifying the best job candidate or you have to decide between two individuals who want to throw a birthday party for you?... this is validating—one of my claims is that we have to make UpTrust good for purely egocentric aims as well, as long as those aims are boundaried such that someone being egocentric on the platform benefits the platform (and vice-versa the platform’s success benefits... i don’t think chatgpt is a good resource for developmental levels. It’s not very good at actually stepping outside of it’s own developmental lens. For examples in the other comment it assumes all the lenses dont like racism or want it to go away, whereas I think you can imagine a... really? I’m surprised in this particular case—I thought it was extremely well established the GPT is biased. Even OpenAI has a help ticket on this. Or the supposedly left-leaning Washington Post article ChatGPT leans liberal, research... Fair enough- llms are not known for their accuracy. But in addition to the other data, I think it’s an interesting point that despite their fine tuning to be as neutral/tame-green-woke as possible, they couldn’t help but begrudgingly admit it (when I asked a few a while...