夜雨聆风Artificial Intelligence
人工智能
We All Use AI. Here’s How to Use It Well
我们都在使用AI工具，如何用好它才是关键
The future belongs to people who can think with AI—without thinking like it.
未来属于那些能与AI一同思考，但不会像AI一样思考的人。
Key points
关键要点
I came to America from Bangladesh at the age of 17, with very little money in my pocket and even less of an idea of what I was walking into. 17岁时，我从孟加拉国来到美国，口袋里钱很少，对自己即将面对什么更是知之甚少。
What I did have —and what I've leaned on for every decision I've made ever since—was a capacity I had developed early in life: the ability to work out what I actually thought, and then to act on it even when no one else agreed with me. 我确实拥有的——并且是我此后做出每一个决定时所依赖的——是我早年就培养出的一种能力：弄清楚自己真实想法的能力，然后即使无人赞同，也依此行动的能力。
There’s a word for that capacity. Judgment. 这种能力有一个词来形容。判断力。
And it’s the thing I'm most worried about losing right now. 而它正是我现在最担心会失去的东西。
The reason I am worried about it is because I use AI every day. 我担心这一点，是因为我每天都在使用AI。
I use it to build applications, develop frameworks, design visual assets, and research what’s happening at the edges of the fields I need to understand. 我用它来构建应用程序、开发框架、设计视觉素材，并研究我需要了解的领域边缘正在发生的事情。
I use it to stress-test arguments before I bring them to my team or to a client. 在将论点提交给我的团队或客户之前，我用它来进行压力测试。
I use it to think through how a message will land before I send it. 在发送信息前，我用它来仔细思考信息会产生何种效果。
In raw output, I'm more productive today than a team of 10 would have been five years ago. 就原始产出而言，我今天比五年前一个十人团队的效率还要高。
And it isn't just speed; the work is objectively better. 而且这不仅仅是速度；工作成果客观上更好了。
Right now, I’m still one of the early adopters. 目前，我仍然是早期采用者之一。
But pretty soon, this will be the reality in every job. 但很快，这将成为每份工作的现实。
Very few people will have a choice about whether or not they use AI, just as few people get to choose whether they use computers or email or the internet today. 很少有人能选择是否使用AI，就像今天很少有人能选择是否使用电脑、电子邮件或互联网一样。
So this isn’t an article about whether to use AI. 所以，这不是一篇关于是否要使用AI的文章。
It’s about how to hold onto your judgment while you do. 而是关于在使用AI时，如何保持你的判断力。
The Core Skill Behind Using AI Well
善用AI背后的核心技能
AI is extraordinarily powerful, but it’s powerful in what we might call a generic way. AI异常强大，但它的强大方式我们可以称之为通用型。
While a generative AI model will be trained on all the insights of all the sciences, all the works of the great artists and the brightest business thinkers, it generally does not and cannot know what matters in your particular situation. 虽然生成式AI模型会基于所有科学的洞见、所有伟大艺术家和最杰出商业思想家的作品进行训练，但它通常不知道，也无法知道在你的特定情况下什么才是重要的。
It does not know what trade-offs you’d accept, what your experience tells you about how something will actually land, or what the right call is given everything you know that the machine doesn’t. 它不知道你会接受哪些权衡取舍，你的经验告诉你某件事实际会产生何种效果，或者根据你所知道而机器不知道的一切，正确的决定是什么。
That knowledge is yours, and using it is what turns AI’s general capability into something that works for you. 这些知识是你的，运用这些知识才能将AI的通用能力转化为对你有效的东西。
Without your judgment, AI gives you fluent but generic output. 没有你的判断力，AI给你的输出流畅但通用。
With it, you get something that couldn’t have come from anyone else. 有了它，你才能得到独一无二、无法从别处获得的东西。
And the combination of AI’s general power and your judgment is far greater than either alone—but only if you’re actively in the conversation, thinking alongside the tool rather than deferring to it. AI的通用能力与你的判断力相结合，其力量远大于任何单独一方——但前提是你必须积极参与对话，与工具一同思考，而不是顺从于它。
This means that using AI well is not fundamentally about writing better prompts or knowing which model to use. 这意味着，善用AI从根本上说，不在于写出更好的提示词或知道使用哪个模型。
Rather, it’s about staying actively engaged with what comes back. 而在于对AI的回应保持积极的参与。
It’s that simple. 就这么简单。
The principle might be simple; the practice is not quite so straightforward. 原则可能很简单；实践却并非如此直截了当。
The sheer ease with which AI models respond to requests and create outputs leads to a phenomenon known as cognitive offloading. AI模型响应请求和创造输出如此轻松，导致了一种被称为认知卸载的现象。
Offloading our mental load is precisely the attraction and the promise of AI: By delegating some tasks we are freed to think more deeply and effectively about other things. 卸载我们的脑力负担正是AI的吸引力和承诺所在：通过委派一些任务，我们得以解放出来，更深入、更有效地思考其他事情。
The risk, though, is that we offload the wrong things—that we outsource our higher judgment about what matters, the ultimate meaning of a piece of work or the creative design. 然而，风险在于我们卸载了错误的东西——我们将关于什么重要、一项工作的最终意义或创意设计的更高层次的判断外包了出去。
If this happens, we stop using AI to support our own thinking and instead begin deferring to the machine. 如果发生这种情况，我们就不是在用AI来支持自己的思考，而是开始顺从机器了。
How to Use AI Well
如何善用AI
These aren’t rules for avoiding using AI. 这些不是避免使用AI的规则。
They’re practices for getting the most out of it—by making sure you’re always the one in the driving seat. 它们是充分利用AI的实践方法——通过确保你始终是那个掌握方向盘的人。
You can’t delegate effectively to an AI model if you aren’t fully in control of the task. 如果你不能完全掌控任务，就无法有效地委派给AI模型。
If you start with a topic you’re interested in and ask the machine what to think about it, you begin by deferring to it. 如果你从一个感兴趣的主题开始，然后问机器对此有何看法，那么你一开始就在顺从它。
Instead, come to it with a developed position you’re willing to defend. 相反，带着你愿意为之辩护的成熟观点去使用它。
That doesn’t mean you shouldn’t ask the AI to challenge your view or help you strengthen it. 这并不意味着你不应该让AI挑战你的观点或帮助你强化它。
But the strategic intent must be yours before you type the first word. 但在你输入第一个字之前，战略意图必须是你自己的。
That’s what makes you the architect of the output rather than turning you into an agent of the AI. 这才能使你成为输出的设计者，而不是变成AI的代理人。
Don’t hand an AI model a blank canvas. 不要给AI模型一张白纸。
Give it a defined problem that includes a sketch of your destination and then use it to help fill in the details. 给它一个定义明确的问题，其中包含你目标的草图，然后用它来帮助填充细节。
That way, the architecture—the requirements, the dimensions, the logic—remain yours. 这样一来，架构——需求、维度、逻辑——仍然掌握在你手中。
If you can’t explain what you’re asking for before you ask, you’re not ready to use the tool. 如果你在提问前无法解释你想要什么，那么你还没有准备好使用这个工具。
The result will be an output that looks finished but that isn’t really yours. 结果将是看起来完成但并非真正属于你的输出。
AI is extraordinarily good at surfacing studies, mapping out a field, and pointing you toward evidence that can confirm or challenge a view. AI非常擅长挖掘研究、勾勒领域轮廓，并指引你找到可以证实或挑战某个观点的证据。
Use it for all of that. 利用它来做所有这些事。
But don’t accept its summary of what the results it digs up mean. 但不要接受它对所挖掘结果意义的总结。
The interpretation—where the evidence leads, what it confirms, where it falls short—has to be your own. 解读——证据指向何方、证实了什么、有何不足——必须是你自己的。
Sometimes we prompt AI because we want help with the processing burden of a task. 有时我们提示AI，是因为我们希望减轻任务的处理负担。
Other times, we reach for it because it offers an easy way around something uncomfortable. 其他时候，我们求助于它，是因为它提供了一种绕过某些不适之事的简便方法。
The discipline is to identify and lean in to the type of difficult that creates friction and to follow where it leads. 这里的修行在于，识别并迎向那种会产生摩擦的困难，并跟随它指引的方向。
This is as true when you are crafting a business strategy, working on an article, or building a product. 无论是制定商业战略、撰写文章还是构建产品，都是如此。
The Human Edge
人类的优势
The practices above are not rules for limiting how or whether you use AI. 上述实践并非限制你如何使用或是否使用AI的规则。
They are the tools I use to ensure I stay in the driver’s seat while I use it, rather than becoming a passenger being moved about by a very capable machine. 它们是我用来确保在使用AI时保持主导地位的工具，而不是变成一个被强大机器载来载去的乘客。
The central skill of the AI age, then, is the willingness to do the driving yourself. 因此，AI时代的核心技能，是愿意自己来驾驶。
To be the architect or orchestrator of the process rather than a storm-tossed ship on a sea of machine reasoning. 成为这个过程的建筑师或指挥家，而不是机器推理海洋上的一叶随波逐流的扁舟。
It means knowing where you’re going before you start out and never letting go of your vision. 这意味着在出发前就知道你要去哪里，并且永不放弃你的愿景。
The most powerful technology ever built will not save anyone who has stopped thinking for themselves. 有史以来最强大的技术，也无法拯救那些停止独立思考的人。
The discipline required for the AI age has two parts: learning to use the systems effectively while refusing to let them use you. AI时代所需的修行包含两部分：学会有效使用这些系统，同时拒绝被它们所用。
About the Author 关于作者
Faisal Hoque is the author of 11 books, including TRANSCEND (2025) and REINVENT (2023). 法萨尔·霍克 是11本书的作者，包括《超越》（2025年）和《重塑》（2023年）。
He is the founder of SHADOKA and NextChapter, among other companies, and serves as an Executive Fellow at IMD Business School in Switzerland. 他是SHADOKA和NextChapter等多家公司的创始人，并担任瑞士IMD商学院的高级研究员。
词汇 (Vocabulary)
短语 (Phrases)
