Why Social Media Feeds You Content You Dislike

算法为何总给你推送讨厌的内容

最新研究显示,社交媒体算法会主动推荐与用户个人价值观相冲突、容易引发愤怒的内容。人们往往倾向于在讨厌的帖子下发表评论,而算法为了最大化用户互动,会将评论互动视为更重要的信号,从而持续推送刺激性内容。这不仅让人陷入愤怒的恶性循环,还加剧了政治极化。对此,研究人员建议通过直接询问用户的价值观,来构建更健康的动态流。
Have you ever wondered why your social media feed is full of posts that make you angry?
你是否曾好奇过,为什么社交媒体动态里总充斥着让你火大的内容?Have you ever wondered 是否曾感到好奇 / social media feed 社交媒体动态流 / make you angry 让你感到愤怒
A new study shows that social media algorithms
一项最新研究显示,社交媒体算法A new study shows 最新研究显示 / social media algorithms 社交媒体算法
actively push content that clashes with your personal values.
会主动推送与你个人价值观相冲突的内容。actively push 主动推送 / clashes with 与……相冲突 / personal values 个人价值观
To study this, psychologists used standard psychological surveys
为了研究这一现象,心理学家采用标准心理学问卷psychologists 心理学家 / standard psychological surveys 标准心理调查
to measure human values.
来衡量人类的价值观。measure human values 衡量人类价值观
Then, they analyzed the feeds of 715 U.S. users on X.
随后,他们分析了 715 名美国 X 用户的动态流。analyzed the feeds 分析了动态流 / U.S. users 美国用户
They found that while people easily give a like to posts they agree with,
他们发现,虽然人们很容易对赞同的帖子点赞,easily give a like 轻易点赞 / posts they agree with 赞同的帖子
they usually write comments on posts they strongly dislike.
但通常会在强烈反感的内容下发表评论。write comments 撰写评论 / strongly dislike 强烈反感
The problem is that the algorithm treats written replies
问题在于,算法将用户留下的文字回复The problem is 问题在于 / treats written replies 将文字回复视作
as a much heavier signal than simple likes.
视作比单纯点赞权重大得多的信号。much heavier signal 权重高得多的信号 / simple likes 单纯的点赞
Because the system is optimized to maximize user engagement,
由于系统旨在最大化用户参与度,optimized to 为……而优化 / maximize user engagement 最大化用户参与度
it quickly serves up more content that provokes you.
它便会迅速推送更多激怒你的内容。quickly serves up 迅速端出/推送 / provokes you 激怒你
As a result, this design traps people in angry feedback loops
结果,这种机制将人们困在愤怒的恶性反馈循环中,traps people 将人困住 / angry feedback loops 愤怒的反馈循环
and increases political polarization.
并加剧了政治极化。increases political polarization 加剧政治极化
To fix this, researchers suggest asking users directly about their values
为了解决这一问题,研究人员建议直接询问用户的价值观,To fix this 为了解决这个问题 / researchers suggest 研究人员建议 / asking users directly 直接询问用户
to build healthier feeds.
以此打造更健康的动态环境。build healthier feeds 构建更健康的动态流

Quiz 🧠

Q1. On social media, people are more likely to write comments on posts they disagree with.
①True
②False
查看答案

✅ True

人们遇到认同的内容通常只是顺手点赞,但面对强烈反对的内容时,更倾向于亲自撰写评论反驳。

People easily like posts they agree with, but strong disagreement makes them want to write replies.

Q2. What does 'provoke' mean in 'posts that provoke you'?
①make you angry
②make you sleepy
③help you relax
④make you laugh
查看答案

✅ make you angry

'provoke' 的意思是激发某种反应,尤其是激起愤怒或烦躁的情绪。

To provoke someone means to cause a strong reaction, especially anger or annoyance.

Q3. Why do social media algorithms show users content that makes them angry?
①To increase user engagement
②To calm people down
③To reduce screen time
④To teach good manners
查看答案

✅ To increase user engagement

令人愤怒的帖子能引发更多的评论互动,从而让用户在应用中停留更长时间。

Angry posts get more written replies, which keeps users active on the app longer.

Dialogue 🎧

Liam: Have you ever wondered why your social media feed is always full of posts that make you angry?
Liam: 你有没有想过,为什么社交媒体动态里总是一堆让人火大的帖子?
Charlotte: I just read a study about that!
Charlotte: 我刚好看了一篇关于这个的研究!
Charlotte: The algorithm actively pushes content that clashes with your personal values.
Charlotte: 算法其实会主动推送跟你个人价值观冲突的内容。
Liam: Wait, really? Why would it purposely try to upset us?
Liam: 等等,真的吗?它为什么要故意惹我们生气啊?
Charlotte: Psychologists used standard surveys to measure human values across 715 U.S. users on X.
Charlotte: 心理学家通过标准问卷调查了 715 名使用 X 的美国用户,衡量了他们的价值观。
Liam: So did people just follow accounts with views they hated?
Liam: 所以是这些人自己关注了观点讨厌的账号吗?
Charlotte: No, their followed accounts actually aligned. The problem is how people interact.
Charlotte: 不是,他们关注的账号其实观念挺一致的。问题出在人们的互动方式上。
Liam: What do you mean? I usually just click like on things I agree with.
Liam: 什么意思?我平时看到赞同的也就是顺手点个赞啊。
Charlotte: We all do! But people write comments mostly when they strongly dislike a post.
Charlotte: 大家都是这样!但人们往往只有在极度反感某个帖子时,才会跑去写评论。
Liam: Oof, guilty. I definitely reply when something makes my blood boil.
Liam: 额,被说中了。看到让人血压飙升的东西,我确实忍不住要回复反驳。
Charlotte: The algorithm treats those written replies as a much heavier signal than simple likes.
Charlotte: 而算法认为这些文字回复的分量比单纯点赞要重得多。
Liam: Because it is designed to maximize engagement, right?
Liam: 因为系统的设计目的就是最大化用户互动,对吧?
Charlotte: Exactly. It serves up more provocative stuff and traps you in angry feedback loops.
Charlotte: 没错。它会喂给你更多刺激性内容,把大家困在愤怒的死循环里。
Liam: That is wild. It explains why political polarization feels so extreme lately.
Liam: 这也太夸张了。难怪最近感觉政治极化变得这么极端。
Charlotte: To fix this, researchers suggest asking users directly about their values to build healthier feeds.
Charlotte: 为了改善这种情况,研究人员建议直接询问用户的价值观,从而打造更健康的动态推荐。
Liam: I would love that. Until then, I am never opening the reply box again.
Liam: 那可太好了。但在改版之前,我发誓再也不打开评论区了。
Charlotte: Good luck with that! Let's see how long you actually last.
Charlotte: 祝你好运哦!我倒要看看你能坚持多久。

在应用中学习这条新闻