Using AI for Job Applications, Honestly

诚实地使用人工智能进行求职申请

One rule settles nearly every case: a model may help you say what is true about you, and it may not decide what is true about you. Drafting is help. Supplying the content of a claim about your own experience is not. 有一条准则适用于几乎所有情况：模型可以帮助你表达关于你自己的事实，但它不能决定什么是关于你的事实。起草是辅助，而提供关于你个人经历的陈述内容则不是。

The line, and why it is there: An application is a set of representations about a person, made by that person, on which somebody else will rely. That is what makes fabrication in one different in kind from fabrication in an essay: there is a party who acts on it, and there are consequences downstream for colleagues, clients and sometimes patients. So the test is not “did a machine touch this”. It is “does the document assert something the applicant does not know to be true”. 这条界限及其存在的原因在于：求职申请是一系列关于个人的陈述，由本人做出，并供他人参考。这就是为什么求职申请中的造假与论文中的造假性质不同：有第三方会根据这些信息采取行动，并会对同事、客户甚至患者产生后续影响。因此，检验标准不是“机器是否参与了”，而是“这份文件是否断言了申请人自己都无法确认真实性的内容”。

A cover letter drafted from your notes and edited by you asserts nothing you did not supply. A cover letter that describes a project you did not run asserts something false regardless of who typed it, and would be equally dishonest written by a friend. 一份根据你的笔记起草并由你编辑的求职信，不会断言任何你未提供的信息。而一份描述了你并未负责过的项目的求职信，无论是由谁打字，都是虚假的陈述，即便由朋友代写也同样是不诚实的。

Case Description

情况说明

Rewriting: fine Rewriting your own bullet points more clearly. Fixing grammar. Translating your industry’s jargon into the target industry’s. Cutting 900 words to 300. Generating ten possible openings so you can choose one. 重写：可以 更清晰地重写你自己的要点。修正语法。将你所在行业的术语翻译成目标行业的术语。将900字删减至300字。生成十个可能的开头供你选择。

Decoding: fine Asking what a job advert is actually asking for, then checking your own experience against that list yourself. 解读：可以 询问招聘广告的真实需求，然后亲自对照该列表检查你自己的经验。

Metrics: not fine Letting it fill in achievements, metrics or responsibilities you have not verified. ‘Increased conversion by 32%’ is a fact about the world; if you do not know the number, it is a fabrication with a number in it. 指标：不可以 让它填补你未核实的成就、指标或职责。“转化率提升32%”是一个客观事实；如果你不知道这个数字，那么它就是一个带有数字的虚构信息。

Motivation: not fine Any statement of motivation you have not read and would not say out loud. ‘I have long admired your work in X’ when you have not is a small lie that is very cheap to expose in an interview. 动机：不可以 任何你未阅读且不会大声说出来的动机陈述。“我一直很钦佩你在X领域的工作”——如果你并非如此，这是一个在面试中极易被拆穿的小谎言。

Assessments: not fine Completing an assessment designed to measure your unaided ability, where the employer has said not to, or where the whole point of the task is the thing you outsourced. 评估：不可以 完成旨在衡量你独立能力的评估任务时，如果雇主明确禁止使用AI，或者任务的核心目的正是你所外包的部分，那么使用AI是不妥的。

What genuinely helps

真正有帮助的做法

The honest uses are also the effective ones, which is convenient. Most applications fail for reasons that have nothing to do with prose quality: they do not address the advert, they bury the relevant experience, or they are written in the vocabulary of the applicant’s current sector rather than the target one. 诚实的使用方式往往也是最有效的方式，这很方便。大多数申请失败的原因与文笔质量无关：它们没有针对招聘广告，掩盖了相关经验，或者使用的是申请人当前行业而非目标行业的词汇。

• Decoding the advert. “List what this advert is actually asking for, separating hard requirements from preferences, and name anything it implies but does not state.” Then you check yourself against the list. This is the highest-value step and it involves no writing at all. 解读招聘广告。 “列出这则广告的真实需求，区分硬性要求和偏好，并指出其中暗示但未明说的内容。” 然后对照列表检查自己。这是价值最高的一步，且完全不需要写作。
• Translation between sectors. A teacher moving into corporate training, a soldier into logistics: the experience transfers and the vocabulary does not. Describing what you did in plain terms and asking how that work is described in the target sector is legitimate and it is genuinely hard to do alone. 行业间翻译。 教师转行企业培训，军人转行物流：经验可以迁移，但词汇不行。用通俗语言描述你的工作，并询问目标行业如何描述这些工作，这是合法的，且单靠个人确实很难完成。
• Ruthless cutting. “Cut this to 250 words, keeping every specific and losing every adjective” produces something better than most people write, because it removes the hedging that a nervous applicant adds. 无情删减。 “将其删减至250字，保留所有具体细节，去掉所有形容词”，这能产生比大多数人写出的更好的内容，因为它去除了紧张的申请人所添加的含糊其辞。
• Finding the gap. “Given this advert and this CV, what will they be unsure about?” It will identify the two-year gap, the sideways move and the missing qualification. You then decide how to address them. 寻找差距。 “根据这则广告和这份简历，他们会对什么感到不确定？” 这将识别出两年的职业空窗期、平级跳槽以及缺失的资格证书。然后你再决定如何应对这些问题。

The CV: tailoring without inventing

简历：定制而非编造

Tailoring is reordering and re-emphasis, not rewriting history. The method that keeps the distinction structural: maintain one long master document containing everything you have actually done, in your own words, with the real numbers where you have them. Every application is generated by selection from that document. 定制是重新排序和强调，而不是改写历史。保持这种区别的结构化方法是：维护一份包含你实际做过的所有事情的完整主文档，使用你自己的语言，并在有数据的地方填入真实数字。每一份申请都应通过从该文档中筛选内容来生成。

Below are two things: a job advert, and my master experience file. Select from the master file only. You may reorder, shorten and re-word. You may not add any responsibility, achievement, tool, figure or date that is not in the file. Produce:

1. The six experience lines most relevant to this advert, in priority order, each rewritten in the advert’s vocabulary.
2. A list headed GAPS: requirements in the advert with nothing matching in the master file. Do not attempt to cover these.
3. A list headed CHECK: anything you have rephrased in a way that might overstate what the master file says. [advert] [master file]

以下是两份材料：一份招聘广告和我的主经验文档。仅从主文档中进行选择。你可以重新排序、缩短和改写。你不得添加任何主文档中不存在的职责、成就、工具、数字或日期。 生成：

1. 与此广告最相关的六行经验，按优先级排序，每行都用广告中的词汇重写。
2. 一个标题为“GAPS（差距）”的列表：广告中要求但主文档中没有匹配项的内容。不要试图掩盖这些。
3. 一个标题为“CHECK（核对）”的列表：任何你改写得可能夸大了主文档原意的内容。 [广告] [主文档]

The GAPS list is the useful output. It tells you what to address directly in the letter, or that this is not your job. The CHECK list exists because rephrasing drifts upwards: “helped coordinate” becomes “led” without anybody deciding it should. GAPS列表是真正有用的输出。它告诉你需要在求职信中直接解决什么问题，或者告诉你这份工作不适合你。CHECK列表的存在是因为改写往往会产生“向上漂移”：比如“协助协调”在无人察觉的情况下变成了“领导”。

Applicant tracking systems parse CVs for keywords, and it is reasonable to use the advert’s own terms where they describe what you did. It is not reasonable to list a tool you have not used because a filter screens for it, and this fails in the most embarrassing possible place, which is a technical question about that tool in the first interview. 申请人跟踪系统（ATS）会解析简历中的关键词，在描述你做过的事情时使用广告中的术语是合理的。但为了通过筛选而列出你从未用过的工具是不合理的，这会在最尴尬的地方——即第一轮面试中关于该工具的技术提问时——导致失败。

The cover letter method

求职信的方法

A generated cover letter reads like a generated cover letter, and recruiters now see hundreds. The failure is not detectable style; it is that the letter contains nothing only this applicant could have written. That is fixable, and the fix is the same one as any other prompt that returns something generic: supply the specifics. 生成的求职信读起来就像是生成的，招聘人员现在每天能看到数百封。失败的原因不在于可检测的文风，而在于信中没有任何内容是只有这位申请人才能写出来的。这是可以修复的，修复方法与任何返回通用内容的提示词一样：提供具体细节。

Write, badly and quickly, three things in your own words: why this job, one thing you have actually done that is relevant, and one genuine question about the role. Give those three things plus the advert to the model and ask for a 250-word letter that uses all three and adds no facts. Read it aloud. Delete any sentence you would not say to the hiring manager in a corridor. This usually removes the opening paragraph entirely. Put back one specific detail that only you know — the thing about the product that annoys you, the reason the sector interests you. This is the sentence that gets a reply. 用你自己的话，快速且粗糙地写下三件事：为什么选择这份工作，你实际做过的一件相关的事情，以及一个关于职位的真实问题。将这三点连同招聘广告一起交给模型，要求它写一封250字的信，使用这三点且不添加任何事实。大声朗读它。删掉任何你不会在走廊里对招聘经理说的话。这通常会删掉整个开头段落。放回一个只有你知道的具体细节——比如关于产品让你感到困扰的地方，或者你对该行业感兴趣的原因。这才是能获得回复的那句话。

Take-home tasks and tests

居家任务与测试

Here the position needs to be plain, because the hedging in most advice is doing no one any favours. If the employer says do not use AI, do not use AI. There is no interpretation under which… 在这里，立场必须明确，因为大多数建议中的含糊其辞对任何人都没有好处。如果雇主说不要使用人工智能，那就不要使用。在任何情况下，都没有理由……