AI本质：企业应如何正确对待人工智能的奇异特性

Ethan Mollick on why employers should treat AI as what it is: weird

伊桑·莫利克认为，雇主应该正视人工智能的本质：怪异的

4月 01, 2026 04:51 上午

ASYSTEM DESIGNEDto predict the most likely next word in a sentence can also write good computer code, offer strategic advice and respond with remarkable empathy to human problems. We don’t fully understand why. And yet the dominant instinct across the corporate world is to treat artificial intelligence as if it were just another piece of enterprise software: slot it into existing processes, assign it Key Performance Indicators and hand it to theITdepartment for management.

一个旨在预测句子中最有可能出现的下一个词的系统，也能编写优秀的计算机代码，提供战略建议，并对人类问题表现出非凡的同理心。我们尚不完全理解其中的原因。然而，企业界的普遍做法却是将人工智能视为另一种企业软件：将其融入现有流程，设定关键绩效指标，然后交给IT部门管理。

This is a profound strategic mistake. Companies are racing to de-weirdAI, and in doing so they are squandering what makes it transformative, turning it into just the latest wave of office automation. To be clear, makingAIeasy to use is important, and building it into familiar tools and workflows is just smart engineering. The mistake is letting those smooth interfaces flatten your understanding of the technology’s possibilities.

这是一个严重的战略失误。企业竞相让人工智能变得“平易近人”，却因此浪费了它变革性的本质，最终将其沦为最新一波的办公自动化工具。诚然，让人工智能易于使用至关重要，将其融入用户熟悉的工具和工作流程也是一种巧妙的工程设计。真正的错误在于，让这些流畅的界面模糊了人们对这项技术潜力的理解。

The urge to de-weird is understandable. Executives are trained to normalise new technologies and fit them into familiar categories. SoAIbecomes a fuzzy-logic processor built into a workflow or a tool that shaves minutes off a task. Normal technology gets normal rollout plans, like requiring that 90% of employees must use a new software package each week.

人们渴望消除对“怪异”事物的抵触情绪是可以理解的。高管们接受的训练是将新技术正常化，并将其归入熟悉的类别。因此，人工智能变成了集成到工作流程中的模糊逻辑处理器，或是能够节省任务时间的工具。而常规技术则采用常规的推广计划，例如要求90%的员工每周都必须使用新的软件包。

But when you set that target forAI, what actually happens? Employees useAIto transcribe meetings, or produce an endless stream of “workslop” in the form of dozens of extra memos or PowerPoints. Treating this technology as another software deployment is like receiving a mysterious alien artefact and immediately using it as a paperweight.

但当你为人工智能设定这样的目标时，实际情况会如何呢？员工们用人工智能来转录会议内容，或者生成源源不断的“工作垃圾”，比如几十份额外的备忘录或PPT。把这项技术当作另一种软件部署来对待，就像收到一件神秘的外星文物却立刻把它当镇纸一样。

The de-weirding impulse produces a second, deeper failure: it leads companies to default towards automation rather than augmentation. When leaders see studies showing productivity gains of 30% fromAI, their instinct is to cut 30% of the workforce. That arithmetic is simple. What is hard, and requires genuine imagination, is asking a different question:

what does it mean to rebuild an organisation around the fact that a single programmer can now write a hundred times more code? What new products become possible? What new markets open up? No vendor can answer those questions for you. No consultant has a playbook (much as they might claim they do). The hard strategic work of reimagining what your organisation could become is precisely the work that de-weirdingAIallows companies to avoid.

这种“去怪异化”的冲动导致了第二个更深层次的失败：它使企业默认选择自动化而非增强。当领导者看到研究表明人工智能可以提高30%的生产力时，他们的本能反应是裁掉30%的员工。这种算术很简单。真正困难且需要真正想象力的是提出另一个问题：

如果一个程序员现在可以编写比以前多一百倍的代码，那么围绕这一事实重建组织意味着什么？哪些新产品成为可能？哪些新市场将会开拓？没有供应商可以回答这些问题。没有咨询顾问能提供现成的方案（尽管他们可能声称自己有）。重新构想组织未来发展方向的艰巨战略工作，恰恰是“去怪异化”人工智能让企业得以逃避的工作。

And there is a natural place where de-weirdedAIgoes to die: theITdepartment. This is not a criticism ofITprofessionals, who do essential work. But in most firms their mandate is to minimise risk. If they could take away your keyboards, they would sleep better at night.

Every creative thing an employee does on a computer leaves the firm potentially vulnerable.AI, by contrast, demands that organisations embrace risk by experimenting wildly, tolerating failure and accepting that nobody yet knows the right way to use these tools. Handing sole control overAIto a department whose core mission is risk elimination is a category error.

而那些经过“去怪异化”的人工智能最终的归宿，自然就是IT部门。这并非对IT专业人员的批评，他们的确在做着至关重要的工作。但在大多数公司，他们的职责是尽可能降低风险。如果他们能收回你的键盘，他们晚上就能睡得更安稳。

员工在电脑上进行的每一项创造性操作，都可能使公司面临潜在的风险。相比之下，人工智能要求组织拥抱风险，大胆尝试，容忍失败，并接受目前尚无人掌握这些工具正确使用方法的现实。将人工智能的全部控制权交给一个以消除风险为核心使命的部门，这无疑是一个范畴错误。

So what should firms do instead? In my work in this area I have come to advocate a three-part model: Leadership, Crowd and Lab.

Leadership means that direction must come from the top. TheCEOand other senior managers cannot delegateAIstrategy to middle management orIT. They must articulate a vision for howAIchanges what the organisationis, not merely how it operates, and they must create incentives that make experimentation safe. And they have to lead by using these systems themselves.

那么企业应该怎么做呢？我在这个领域的研究中提倡一种三部分组成的模式：领导力、群体和实验室。

领导力意味着方向必须来自高层。首席执行官和其他高级管理人员不能将人工智能战略委托给中层管理人员或IT部门。他们必须阐明人工智能将如何改变组织的本质，而不仅仅是改变组织的运作方式，并且他们必须制定激励机制，确保实验的安全进行。此外，他们还必须以身作则，亲自使用这些系统。

If they do, they can inspire the Crowd, a company’s employees, who, when given access toAItools and genuine permission to experiment, will figure out use cases not even theAIfirms expected. SinceAIis most effective in the hands of experts, the Crowd is where the best ideas come from.

如果他们做到了，就能激励“大众”（即公司员工），当员工们获得人工智能工具的使用权限和真正的实验许可后，他们会发现连人工智能公司都意想不到的应用场景。由于人工智能在专家手中才能发挥最大效用，因此，最佳创意往往来自“大众”。

Those ideas then go to the Lab, where a team of technical and non-technical employees work on generativeAIfull-time. These are people whose job is to push boundaries, develop new workflows and feed discoveries back into the organisation. I am shocked by how many large companies still lack even this. Without it, they have no mechanism for learning whatAIcan actually do for them. They are flying blind, relying on vendor demos and conference keynotes instead of building institutional knowledge.

这些想法随后会被送到实验室，由一支技术和非技术人员组成的团队全职从事生成式人工智能的研究。他们的工作是突破界限，开发新的工作流程，并将研究成果反馈给公司。令我震惊的是，许多大型公司甚至连这一点都缺乏。没有它，他们就无法了解人工智能究竟能为他们做些什么。他们就像盲人摸象，依赖供应商的演示和会议主题演讲，而不是建立机构知识体系。

There is one more problem that de-weirding creates, and it may be the most consequential. When companies fail to create the right incentives, employees respond rationally: they hide theirAIuse. Some fear punishment. Some do not trust that productivity gains will be shared with them rather than captured by the firm. Some quietly work 90% less and see no reason to volunteer that information. The result is an enormous information gap. Managers cannot see the true impactAIis already having inside their own organisations, which makes it even harder to develop a real strategy.

去怪异化还会带来另一个问题，而且这个问题或许最为严重。当公司未能制定合适的激励机制时，员工会做出理性的反应：他们会隐瞒自己使用人工智能的情况。有些人害怕受到惩罚。有些人不相信公司会将生产力提升的成果与他们分享，而是据为己有。有些人悄悄地减少了90%的工作时间，却觉得没有必要主动告知。其结果是造成了巨大的信息鸿沟。管理者无法看到人工智能在其组织内部已经产生的真正影响，这使得制定切实有效的战略变得更加困难。

Resisting the de-weirding trend does not guaranteeAIwill go well. There is no default good outcome. But a failure to seeAIfor what it is—a profoundly odd, risky and powerful technology—will guarantee bad ones. Firms that sand downAI’s strange edges will veer towards automation and layoffs, as that is all they can see. Those willing to confront the technology head-on can find something far more interesting, including ways to help make their people, and their organisations, capable of things that were impossible a year ago and will be impossible to predict a year from now. Nobody knows exactly where this is going. But you don’t navigate strange territory by pretending your old maps will work. ■

抵制“去怪异化”的趋势并不能保证人工智能会一帆风顺。没有默认的完美结果。但如果不能认清人工智能的本质——一种极其奇特、充满风险且威力强大的技术——则必然会导致糟糕的结果。那些试图掩盖人工智能怪异之处的公司，最终只会走向自动化和裁员，因为他们只能看到这些。而那些愿意直面这项技术的公司，则能发现更有趣的东西，包括如何帮助他们的员工和组织实现一年前不可能、一年后也无法预测的事情。没有人能确切地知道人工智能的未来走向。但你不能指望用旧有的思维模式来应对未知的领域。

PS:

Ethan Mollick is an associate professor at the Wharton School and the author of “Co-Intelligence: Living and Working withAI”.

Ethan Mollick 是沃顿商学院的副教授，也是《协同智能：与人工智能共同生活和工作》一书的作者。

To share with friends

北京打工人自传《我在北京送快递》 爆火畅销海外

汤姆·斯坦迪奇预测2026年十大趋势

职业阶梯最底层会消失吗--人类会不会被AI取代

当代儿童被人工智能重塑的童年

水军？AI? 人工智能的其他后果