摘要：AI can generate false yet authoritative-sounding information.人工智能可以生成虚假但听起来权威的信息。Emotional manipulation accelerates the spread of

Thinking Clearly in the AI Era

AI时代的清晰思考

Why evidence-based thinking will always matter.

为什么循证思维永远很重要。

Posted September 2, 2025 | Reviewed by Abigail Fagan

发布于 2025 年 9 月 2 日 |阿比盖尔·费根 （Abigail Fagan） 点评

AI can generate false yet authoritative-sounding information.

人工智能可以生成虚假但听起来权威的信息。

Emotional manipulation accelerates the spread of misinformation.

情绪纵加速了错误信息的传播。

Evidence-based thinking is the best tool against misinformation.

循证思维是对抗错误信息的最佳工具。

Today, information spreads faster than ever. Social media, marketing campaigns, and AI tools provide rapid access to knowledge, but not all of it is accurate.

如今，信息传播速度比以往任何时候都快。社交媒体、营销活动和人工智能工具提供了快速获取知识的机会，但并非所有知识都是准确的。

AI chatbots such as ChatGPT-5 and Claude are powerful, convincing, and increasingly used for professional, educational, and personal guidance. Yet they can also generate fake and incorrect information, statements that sound plausible but are entirely false.

ChatGPT-5 和 Claude 等人工智能聊天机器人功能强大、令人信服，并且越来越多地用于专业、教育和个人指导。然而，它们也可能产生虚假和不正确的信息，这些陈述听起来似是而非，但完全是错误的。

People often trust AI because it is authoritative, articulate, and seemingly objective. But confident-sounding information can still be completely wrong. The result is an illusion of credibility.

人们通常信任人工智能，因为它权威、清晰且看似客观。但听起来自信的信息仍然可能完全错误。结果是一种可信的错觉。

The Authority Bias

权威偏见

Large language models such as ChatGPT can produce authoritative-sounding responses, often complete with professional-looking citations. Yet many of these references are fabricated.

ChatGPT 等大型语言模型可以产生听起来权威的响应，通常配有具有专业外观的引文。然而，其中许多参考资料都是捏造的。

Gravel and colleagues, for instance, evaluated 59 references provided by ChatGPT and found that almost two-thirds were fabricated, despite appearing legitimate with plausible authors, journal titles, and DOI numbers (Gravel et al., 2023).

例如，Gravel 及其同事评估了 ChatGPT 提供的 59 份参考文献，发现近三分之二是捏造的，尽管作者、期刊标题和 DOI 编号看起来是合法的（Gravel 等人，2023 年）。

In medicine, this can be especially dangerous: inaccurate references, poor descriptions of conditions, and fabricated case details have already misled students and professionals.

在医学领域，这可能尤其危险：不准确的参考文献、对病情的糟糕描述和捏造的病例细节已经误导了学生和专业人士。

These errors are not intentional deception but the byproduct of how AI generates text, drawing patterns from training data without verifying facts or consulting real databases. The result is content that looks trustworthy but may contain serious inaccuracies with real-world consequences.

这些错误不是故意欺骗，而是人工智能生成文本、从训练数据中绘制模式而不验证事实或查阅真实数据库的副产品。结果是内容看起来值得信赖，但可能包含严重的不准确之处，从而产生现实世界的后果。

The legal world has also seen this problem. AI has invented fake case law and used fictional information. It has generated legal cases that sound real, complete with credible details, which lawyers have then used in court filings. In pharmacology, AI chatbots have recommended prescription medications that do not exist, citing fake studies. In history, AI has blended fact with fiction in ways that sound entirely plausible.

法律界也看到了这个问题。人工智能发明了虚假判例法并使用了虚构信息。它产生了听起来真实的法律案件，并附有可信的细节，然后律师在法庭文件中使用了这些案件。在药理学方面，人工智能聊天机器人引用了虚假研究，推荐了不存在的处方药。在历史上，人工智能以听起来完全合理的方式将事实与虚构融合在一起。

Across all these domains, the same pattern emerges AI’s confident-looking outputs make false information persuasive. This illustrates authority bias, our inclination to believe information from sources that appear expert, even without independent verification.

在所有这些领域中，出现了相同的模式：人工智能看起来自信的输出使虚假信息具有说服力。这说明了权威偏见，即我们倾向于相信来自看似专业的来源的信息，即使没有独立验证。

Synthetic Empathy and Emotional Pull

综合同理心和情感吸引力

Authority bias is not the only lever of influence. People also respond strongly to AI’s recently developed synthetic empathy, its ability to mirror human emotions, express understanding, or offer compliments. This sycophantic, personality-like behavior may make users more likely to accept AI statements as true, even when they are false.

权威偏见并不是唯一的影响力杠杆。人们还对人工智能最近开发的合成同理心、反映人类情感、表达理解或赞美的能力做出了强烈反应。这种阿谀奉承的、类似个性的行为可能会使用户更有可能接受人工智能的陈述是真实的，即使它们是虚假的。

Our Feelings Can Be Hijacked

我们的感情可能会被劫持

Our feelings can be just as easily manipulated as our acceptance of false information. A well-known example is the soldier reunion ads, in which soldiers were shown reuniting with their families in surprise homecomings. While people felt genuine empathy and emotion, the ads were later criticized for exploiting those feelings to solicit donations and support without being transparent about their true purposes.

我们的感受就像我们接受虚假信息一样容易纵。一个著名的例子是士兵团聚广告，其中士兵们在出人意料的回家中与家人团聚。虽然人们感受到了真正的同理心和情感，但这些广告后来因利用这些感觉来募捐和支持而没有透明地说明其真实目的而受到批评。

Our empathy can be hijacked before we even realize it. Emotional manipulation is especially effective because humans evolved to respond to social cues and narratives. When empathy, fear, or anger is triggered, we are less likely to pause, reflect, or verify information. This makes us vulnerable to disinformation, marketing ploys, and misleading AI outputs that combine emotional language with persuasive authority.

我们的同理心可能会在我们意识到之前被劫持。情绪纵特别有效，因为人类进化为对社会线索和叙事做出反应。当同理心、恐惧或愤怒被触发时，我们不太可能停下来、反思或验证信息。这使得我们容易受到虚假信息、营销策略和误导性人工智能输出的影响，这些输出将情感语言与说服权威相结合。

Why Thinking Critically Matters

为什么批判性思考很重要

Together, AI misinformation and emotional manipulation create a perfect storm: We encounter information that looks credible, appears convincing, and pulls at our emotions, all before we have a chance to reflect.

人工智能错误信息和情绪纵共同创造了一场完美风暴：我们遇到了看起来可信、看起来令人信服并拉动我们情绪的信息，所有这些都在我们有机会反思之前。

AI can mimic the authority of science, and it can also exploit human empathy to persuade and manipulate. Our cognitive biases, including confirmation bias and authority bias, and our susceptibility to emotional influence exacerbate the problem.

人工智能可以模仿科学的权威，也可以利用人类的同理心进行说服和纵。我们的认知偏差，包括确认偏差和权威偏差，以及我们对情绪影响的敏感性加剧了这个问题。

This is where evidence-based thinking provides exactly the toolkit we need. It is the best defense against both AI errors and emotional manipulation. It helps us resist the allure of authority and the pull of manipulated emotional responses.

这就是循证思维提供我们需要的工具包的地方。它是针对人工智能错误和情绪纵的最佳防御措施。它帮助我们抵制权威的诱惑和纵的情绪反应的牵引。

Scientific, evidence-based thinking provides a framework for thinking critically without imposing conclusions. Unlike approaches that teach us what to believe, this method teaches us how to evaluate, question, and reason.

科学的、基于证据的思维提供了一个批判性思考的框架，而无需强加结论。与教我们相信什么的方法不同，这种方法教我们如何评估、质疑和推理。

Evidence-Based Thinking Skills We Can Use Today

我们今天可以使用的循证思维技能

Applying scientific, evidence-based thinking is straightforward. It includes the following practices:

应用科学、基于证据的思维很简单。它包括以下做法：

Evaluating AI outputs: Before acting on AI-generated information, ask: What evidence supports this? Are there credible sources? Could this be fabricated?

评估人工智能输出：在对人工智能生成的信息采取行动之前，先问：什么证据支持这一点？有可靠的来源吗？这会是捏造的吗？

Analyzing emotionally charged content: Notice if empathy, fear, or outrage is being triggered. Pause and examine whether the emotional impact is influencing your judgment.

分析充满情绪的内容：注意是否触发了同理心、恐惧或愤怒。停下来检查情绪影响是否影响了你的判断。

Navigating news and social media: When encountering a striking headline, check for multiple sources. Consider whether it confirms preexisting beliefs and whether emotions are guiding your reactions.

浏览新闻和社交媒体：当遇到引人注目的标题时，请检查多个来源。考虑一下它是否证实了预先存在的信念，以及情绪是否在指导你的反应。

Decision-making in professional contexts: Use hypothesis-driven thinking to assess claims, proposals, or data before committing to action.

专业环境中的决策：在采取行动之前，使用假设驱动的思维来评估主张、提案或数据。

article continues after advertisement

广告后文章继续

In sum, evidence-based thinking requires us to pause, question, and evaluate.

总之，循证思维需要我们停下来、质疑和评估。

Why Evidence-Based Thinking Matters More Than Ever

为什么循证思维比以往任何时候都更重要

AI tools, social media, and emotionally charged messaging are accelerating the spread of false or misleading information. Decisions about health, finances, civic engagement, and personal relationships increasingly hinge on the quality of the information we trust. Without methods to critically assess what we see, we risk making choices based on errors or manipulation.

人工智能工具、社交媒体和充满情感的信息正在加速虚假或误导性信息的传播。有关健康、财务、公民参与和人际关系的决策越来越取决于我们信任的信息的质量。如果没有批判性地评估我们所看到的东西的方法，我们就有可能根据错误或纵做出选择。

Learning how to think rather than what to think is not just an academic exercise; it is a practical necessity. Evidence-based thinking provides a roadmap for navigating a world where authority and emotion can be weaponized. It can help us become more thoughtful, reflective, and resilient in the face of misinformation. By adopting evidence-based thinking, we can cultivate habits of careful reasoning, balanced evaluation, and awareness of bias.

学习如何思考而不是思考什么不仅仅是一项学术练习;这是实际需要的。循证思维为驾驭一个权威和情感可以被武器化的世界提供了路线图。它可以帮助我们在面对错误信息时变得更加深思熟虑、反思和有弹性。通过采用循证思维，我们可以培养仔细推理、平衡评估和偏见意识的习惯。

Ultimately, the goal of evidence-based thinking is to empower us to ask the right questions, evaluate evidence, and make informed decisions, skills that are more important than ever in today’s rapidly changing, misinformation-rich environment. Authority and emotion are powerful, but they are not invincible forces in shaping belief.

最终，循证思维的目标是使我们能够提出正确的问题、评估证据并做出明智的决定，这些技能在当今快速变化、错误信息丰富的环境中比以往任何时候都更加重要。权威和情感是强大的，但它们并不是塑造信念的无敌力量

来源：左右图史