TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Gemini's AI Image Generator Accused of Being 'Anti-White'

10 点作者 giuliomagnifico超过 1 年前

2 条评论

eadmund超过 1 年前
The Ministry of Truth in Orwell’s 1984 would have loved this sort of thing. Why go to the work of manually rewriting history when you can just generate a new one on demand?<p>It is disturbing how much of our sense of reality is based on media we absorb rather than experiences we have. I imagine that we all agree that media depicting solely minorities as criminals has a negative effect, but I think that applies to <i>all</i> false representation. Most police aren’t actually thugs, and media portraying them thus may actually lead them to think of themselves that way. Soldiers aren’t automata who blindly follow orders. A 90-pound teenager probably <i>isn’t</i> going to be able to defeat a 220-pound brawler. And on &amp; on. But if one subsists on a diet of lying media and AI for long enough, one might believe these things.<p>Generative AI should strive to be actually unbiased. That means it should not skew numbers in <i>either</i> direction, for <i>anyone</i>.<p>On another generative-AI note, it is remarkably difficult to get ChatGPT to use the masculine pronoun to refer to unknown persons. Including ‘prefer masculine pronouns’ tends to result in sentences such as ‘Hey guys, this bad boy sure is great, man’ as opposed to just ‘this item is great’ — and can still end up with superfluous ‘they’ anyway!
lostemptations5超过 1 年前
No one will ever be satisfied-- let&#x27;s just let people create what they want. They can describe it in the prompt.
评论 #39469207 未加载