By 2026, artificial intelligence is no longer a novelty on college campuses; it is a daily utility, as ordinary as a library card or a vending machine. Yet the ease with which students now summon an essay draft, a coding tutorial, or a second opinionfrom a chatbot has exposed a deeper split.

On one side are the "accelerators", who treat AI as a turbocharger for their own curiosity. They use it to brainstorm essay topics, to translate dense academic papers into plain English, or to run through mock interviews the night before a big career fair. For them, the tool is only as good as the questions they know how to ask.

On the other side are the "avoiders", a smaller but vocal group who worry that relying on AI is quietly eroding the very skills a university is supposed to teach — the patience to read a long chapter, the confidence to draft a clumsy first paragraph, the ability to argue without a safety net. Professors, many of whom were raised on a no-calculator arithmetic, share this unease.

The most interesting students, however, occupy the middle. They neither worship the tool nor reject it outright. Instead, they have developed a quiet habit of editing the machine: checking its citations, rewriting its bland sentences, and treating its output not as an answer but as a starting point. In doing so, they may have landed on the most honest definition of digital literacy in the age of AI.

The divide is unlikely to close soon. What is clear is that the next generation of graduates will be judged less by what they know than by how they judge what they are told — by the machines, and by one another.

到 2026 年，人工智能在大学校园里早已不是什么新鲜事；它已经像借书卡、像楼下的自动售货机一样，成了日常工具。然而，当学生们随手就能从聊天机器人那里召来一篇论文初稿、一段编程教程、或是一份"第二意见"时，一种更深层的分歧也随之浮出水面。

一边是"加速器派"——他们把 AI 当成好奇心的涡轮增压器。用它来头脑风暴论文选题，把晦涩的学术论文翻译成大白话，或者在大型招聘会前一晚反复模拟面试。对他们而言，工具好不好用，完全取决于你会不会提问。

另一边是"回避派"——人数不多，但声音响亮。他们担心，依赖 AI 正在悄悄侵蚀一所大学本应教会你的那些能力：读完一整章书的耐心、笨拙地写下第一段的勇气、以及在没有安全网的情况下独立论证的能力。许多教授也是在"不许用计算器"的年代长大的，他们同样心怀不安。

但最有意思的，是站在中间的那批学生。他们既不盲目崇拜工具，也不彻底拒绝它。相反，他们养成了一种安静的习惯——"编辑机器"：核查它给出的引用，重写它干巴巴的句子，把它的输出不当作答案，而当作起点。正是在这样做的过程中，他们或许找到了 AI 时代数字素养最诚实的定义。

这道分水岭短期内不会弥合。但有一点已经清楚：下一代毕业生将被评判的，不再是"你知道什么"，而是"你如何判断别人告诉你的东西"——无论是机器告诉你的，还是彼此告诉的。