✦   ✦   ✦

breaking news

News, read through The Primal Race
← 全部评论 · all commentary

AI 教你找谁替你干活AI Will Find Someone to Do It for You

科技 文化层 · 结构层 404 Media ↗ 2026-10-31 § 链接
新闻摘要404 Media 的 Jason 撰文讲述 ChatGPT 如何让他成为人们随机技术问题的「技术支持」。文章发表后,他收到更多由 AI 生成的求助邮件,其中不少还指导求助者同时联系多位记者、律师或学术研究者。
The news404 Media's Jason wrote about ChatGPT turning him into tech support for people's random problems. After the article, he received more AI-generated emails asking for help, many telling people to contact multiple journalists, lawyers, or academic researchers.
工具替你写了求助信,代价还是别人付。
The tool writes the plea; someone else still pays.

ChatGPT 把一群人变成了另一群人的技术支持,这件事的重点不是 AI 会写邮件,是它教会人把成本转嫁给谁。

404 Media 的 Jason 写过一个故事:ChatGPT 把他变成了别人随机技术问题的客服。文章发出来之后,他收到的 AI 生成求助邮件更多了。邮件里的人不是自己解决问题,是让 AI 告诉他们该找谁解决——AI 给了记者的联系方式,也给了联系律师和学者的建议,还附上操作步骤。

这是表达被武器化的一个很小的切片,小到几乎无害,但机制完整:AI 在这里不是工具,是认知入口。它替使用者决定了「这件事该由谁承担」,而这个决定从来不需要经过被指定的人同意。写信的人省下了搜索、判断、试错的时间,Jason 们收下了这些时间——以邮件的形态,以又一份无薪劳动的形式。

值得注意的是 AI 建议的名单:记者、律师、学术研究者。全是靠处理他人问题吃饭的人,也全是时间被默认可以征用的人。AI 没有发明这种预期,它只是把这种预期自动化了,还贴心地附上了联系方式。

真正的问题不在垃圾邮件。在于一个人遇到问题时,第一反应已经不是「我该怎么弄明白」,而是「谁该替我弄明白」——而 AI 恰好很擅长回答第二个问题。

Jason 说希望文章里写清楚了,这不是 404 Media 一家的事。他写清楚了。但收到邮件的那些人显然没读到那一段,或者读到了,然后让 ChatGPT 帮他们写了一封新的。

ChatGPT turned a bunch of people into tech support for a different bunch of people. The story is not that AI can write emails. It's that AI taught people whom to offload the cost onto.

Jason at 404 Media wrote about ChatGPT turning him into customer service for strangers' random problems. After the piece ran, he got more AI-generated requests, not fewer. The people writing them weren't solving anything themselves. They were asking AI to tell them who should solve it. The AI supplied journalists' contact info, suggested lawyers and academic researchers, and included instructions on how to reach out.

This is a very small slice of expression weaponized, small enough to look harmless, but the mechanism is intact: the AI is not a tool here, it's a cognitive entry point. It decides, on the user's behalf, who should bear this cost, and that decision never passes through the person being assigned. The sender saves the time of searching, judging, trying. Jason collects that time, in the form of emails, in the form of another unpaid task.

Note the list the AI recommends: journalists, lawyers, academic researchers. All people who eat by handling other people's problems. All people whose time is assumed to be available for requisition. AI didn't invent that assumption. It automated it, and helpfully attached contact details.

The real issue isn't spam. It's that when a person hits a problem, the first reflex is no longer "how do I figure this out" but "who should figure this out for me"—and AI happens to be very good at answering the second question.

Jason said he hopes the article made clear this isn't just a 404 Media thing. He made it clear. The people sending the emails evidently didn't read that part, or read it and had ChatGPT draft a new one.