摘要：卡夫卡的这句话，揭示了写作的本质：孤独、内省，以及与自我深处的对话。然而，在人工智能迅速渗透的今天，写作是否仍是那场独自的精神探险？还是正如一些人担忧的那样，它正在被算法改写为另一种轻松而热闹的“群聊”？
写作与AI：过程与结果的再思考
“Writing is utter solitude, the descent into the cold abyss of oneself.”— Franz Kafka
写作是彻底的孤独，是坠入自身冰冷深渊。
弗朗茨·卡夫卡
（Franz Kafka，1883–1924）
卡夫卡的这句话，揭示了写作的本质：孤独、内省，以及与自我深处的对话。然而，在人工智能迅速渗透的今天，写作是否仍是那场独自的精神探险？还是正如一些人担忧的那样，它正在被算法改写为另一种轻松而热闹的“群聊”？
在《纽约时报》“伦理学家”专栏中，就有这样一封来信：
一位历史学会志愿者提到，他们每年都会向两名高中生颁发1000美元奖学金，条件是提交关于历史遗址的作文。今年学生作文的质量大幅提升，但直到宣布结果后，他们才意识到其中一位获奖者几乎肯定使用了人工智能。提问者困惑：我们该怎么办？要不要让她退还奖金？
专栏作者回应道：人工智能工具正在让作文竞赛、奖学金评审乃至学校考试的公平性变得棘手。因为独立调研和写作，向来是学生学习知识、形成独立见解的重要过程。而如果作文仅仅成了机器生成、人工修饰的产物，那么教育所强调的“思考”环节就被削弱了。
他提醒道，这更像是一个“警醒信号”而不是“犯罪现场”。与其逼迫学生认错，不如明确规则，告诉他们哪些帮助是越界的。同时，评审形式也该更新，比如增加现场写作、口头汇报，以确保学生真正理解并表达自己的想法。
与此呼应，国内最近也有类似的争论。2025年1月，腾讯推出的“朱雀AI检测”工具升级后，许多自媒体文章的AI“含量”被迅速检出，引发轩然大波。有人担心自己的原创被误判为AI生成，有人则恐惧使用AI后“原形毕露”。一时之间，写作者与读者都在焦虑：什么才算“真实的写作”？
但冷静想想，借助外力从来不是新鲜事。几十年前，孩子写作文时请父母修改，是“作弊”吗？今天用拼写检查、语法校对工具，又算违规吗？现实生活中，家长帮忙做小报、改作文，早已是常态。关键不在于是否使用工具，而在于：学习的核心过程有没有发生？学生是否真正思考、掌握并表达了自己的观点？
因此，写作的价值不在于文章是否百分之百“人手码出”，而在于过程中是否经历了研究、思考与表达。AI可以成为辅助工具，帮助理清思路、完善文稿，但学习与成长的核心仍必须由学生亲自完成。教育的目标，应是让他们理解：工具可以用，但思想要靠自己。
卡夫卡所说的“坠入冰冷深渊”，或许在今天已被算法的“温暖群聊”部分取代。但真正的写作与学习，依然需要那份孤独与思辨。唯有如此，我们才能让AI成为助力思维的伙伴，而不是割裂学习的屏障。
An Essay Contest Winner Used A.I. Should She Return the $1,000 Award?
I volunteer with our local historical society, which awards a $1,000 scholarship each year to two high school students who submit essays about a meaningful experience with a historical site. This year, our committee noticed a huge improvement in the quality of the students’ essays, and only after announcing the winners did we realize that one of them, along with other students, had almost certainly used artificial intelligence. What to do? I think our teacher liaison should be told, because A.I. is such a challenge for schools. I also feel that this winner should be confronted. If we are right, that might lead her to confess her dishonesty and return the award. — Name Withheld
From the Ethicist:
The rise of A.I. chatbots — and now Microsoft Word’s Copilot feature — is making it harder to use essays to assess candidates for scholarships, grades and jobs. In fact, because much of the workplace will soon involve prompting machines and editing their output, traditional essay writing may no longer be seen as an essential skill. One big worry about all this is that researching and drafting a paper on their own has long been one of the ways students actually learn material and form their own views about it.
Your society’s prize is no doubt meant to reward that process, not just a polished final product. And yet unless your historical society explicitly barred A.I. use, the winner might not have thought she was doing anything wrong. Perhaps she provided information about a historical site and her experience of it and then refined what the machine produced.
Before generative A.I., of course, students could already run their work through digital spelling, grammar and style checkers. Earlier still, they might have had a parent or an older sibling make improvements. Now services like Grammarly boast that they help users construct an essay interactively. “Writing is utter solitude,” Kafka is said to have lamented, “the descent into the cold abyss of oneself.” For students today, it may feel more like a cozy group chat with an algorithm.
What happened this year should be taken as a wake-up call, rather than a crime scene. If you’re going to evaluate essays written without supervision, it’s wise to talk to the students and confirm that they actually understand the ideas in their papers. For now, let the teacher liaison know that many of the submissions appeared to rely on generative A.I. But I wouldn’t push for a confession from the winner. What matters is letting students know what kinds of assistance are off-limits and why producing work on their own is (as you and I believe) still worth doing.
If the contest continues in its current form, however, you may simply be rewarding those crafty enough to conceal their silicon Cyranos. I suspect that it’s time to rethink the format. That might mean shifting to supervised writing sessions or oral presentations about a site or something else entirely. The goal would still be to encourage young people to explore the past — to research, reflect and communicate. But to do that, the contest itself may need a rewrite.
竞赛获奖作文用了人工智能，她是否该退还1000美元奖金？
我在当地历史学会做志愿者。学会每年都会向两名高中生颁发奖学金，各1000美元，条件是提交一篇关于某个历史遗址的文章，要能分享他们有意义的经历。我们委员会注意到今年学生作文的质量有了大幅提升。直到公布获奖者之后我们才意识到，其中一位几乎可以肯定使用了人工智能（其他几位或多或少也用了）。这就难办了。我认为我们应该把情况告诉对接的老师，毕竟人工智能对学校来说是个很大的挑战。我也觉得应该直接与这位获奖者当面聊聊。如果我们判断无误，或许她会承认自己的不实行为，并退还奖金。——匿名提问
伦理学家的回应：
人工智能机器人以及如今微软 Word 的 Copilot 功能，正在让奖学金评定、成绩认可乃至求职信的考察变得更加棘手。事实上，职场中越来越多的工作都可以通过向机器下达指令并由人工智能输出结果搞定，传统的写作可能不再被视为一种必备技能。令人担忧的是，长期以来独立调研和论文写作一直是学生真正学习、形成独立观点的途径之一。
贵学会的奖项显然是为了奖励这一过程，而不仅仅是一篇经过打磨的成品。然而，除非你们在规则中明确禁止使用人工智能，否则获奖者未必觉得自己做错了什么。她或许提供了历史遗址及相关经历的信息，再由机器人润色而成。
在生成式人工智能出现之前，学生们早就可以用拼写、语法和风格检查工具修改作文（你会禁止别人使用文本编辑软件中的“拼写检查”功能吗？）。在更早的时候，他们甚至会请父母或兄长帮忙改进。如今，像 Grammarly 这样的服务，声称能帮助用户用互动的方式完成写作。“写作就是彻底的孤独，”据说卡夫卡曾感叹过，“是坠入自我的冰冷深渊。”而对当今的学生来说，这更像是跟算法之间面对面的一场暖意盎然的群聊。
今年贵学会奖学金评定过程中发生的情况，与其说是一起“犯罪现场”，不如说是一个“警醒信号”。如果要在无人监督的情况下评估作文，明智的做法是和学生谈一谈，确认他们是否真正理解自己文章中的思想。眼下最好告诉联络老师，很多投稿显然依赖了生成式人工智能，这就足够了。但我不会建议你逼迫获奖者认罪，承认自己不诚实甚至承认造假。关键在于让学生明白自己获得的帮助中，哪些是越界的，以及为什么坚持独立完成作品（如你我所相信的那样）是值得的。◾
来源：左右图史