Transcript | 节目原稿
嗨，你好，Bonjour [bɔ̃.ʒuʁ]! 这里是 Jojo‘s Pod。今天我们来聊聊 AI，有人觉得它是巨大威胁，也有人觉得它能解放我们人类。
Hi there, Bonjour [bɔ̃.ʒuʁ]! This is Jojo’s Pod. Today we’ll talk about AI — some see it as a great threat, while others believe it can liberate humanity.
chatGPT刚刚出现时，我也迫不及待地注册使用；那时我把它当作百科全书，辅助我展开对问题的探究。
When ChatGPT first came out, I quickly signed up. Back then, I treated it like an encyclopedia to help me explore questions.
最初令人感到兴奋，它的确可以胜任一些主题的解释，但后来我发现它有不小的幻觉，关键问题还是需要我自己一步步去搜索查询，并且，它的理解也会有偏差，再然后，登录还总会时不时出问题，我就渐渐不太用它了。
At first it was exciting — it could handle some topics well. But soon I noticed its “hallucinations.” On key questions, I still had to research step by step myself, and its interpretations sometimes went astray. Plus, login kept failing now and then, so I used it less and less.
直到今年我再次用回它，明显觉得它进化了好多。
Only this year, when I returned to it, did I feel how much it had evolved — more accurate, more useful, and more woven into daily life.
先是发现它在提高阅读效率方面很有用。我那时在读几本比较难懂的哲学书，需要对照英、法、德原文与不同译本，chatGPT可以给出准确的译文，列出术语表，并详细解释每一个重要概念，并且，它可以延展你的思考，帮你拓展一些其他的相关概念与哲学观点。
I first found it useful in reading. While studying some difficult philosophy books, I needed to compare English, French, and German originals with different translations. ChatGPT could provide accurate translations, glossary lists, detailed explanations of concepts, and even extend my thinking by bringing in related ideas.
当然它还是会有幻觉，如果你没有给它提供原文，它有时会凭空拼凑出一段看起来像原文的东西，虽然意思近似。
Of course, it still hallucinates: if I don’t feed it the original text, sometimes it fabricates a passage that looks like the source but isn’t — though the meaning may be close.
所以使用时一定要有判断，但尽管如此，作为阅读助手，必须说它已经超级棒。
So you need judgment when using it. But still, as a reading assistant, I must say: it’s fantastic.
随后我也在日常生活中发现它的用处。因为在阅读上它对我帮助很大，也因为我认真考察过不同 AI 公司在伦理上的承诺和实践，这让我对 ChatGPT 的信任逐渐增加，也才让我敢把一些生活相关的信息告诉它。
Later, I discovered its role in daily life. Because it had helped me in reading, and because I had studied the ethical commitments of different AI companies, my trust in ChatGPT grew — enough for me to share some personal information with it.
故事要从一块手背上的皮疹说起。大概是冬天没注意保暖，起了两个小红点，我以为是冻疮，随手抹点凡士林，想着会慢慢好。结果到了初夏，情况反而变严重，我又自作聪明一通瞎操作，乱用了一些外用药——百多邦、红霉素、薄荷膏、甚至贴了创可贴——几天之内反而恶化，变成一块约一平方厘米、有渗液的湿疹。那只手几乎不能碰水，生活很受影响。
The story begins with a rash on my hand. In winter, I neglected warmth, got two small red spots, thought they were frostbite, and casually applied Vaseline. By summer, it worsened. I foolishly used random ointments — antibiotic creams, menthol balm, even a bandage. Within days it became an oozing eczema patch about a square centimeter wide. That hand could hardly touch water, making life very inconvenient.
我忽然想起，或许可以问问 AI。于是我把照片和描述发给它。它的反应很直接：先严肃建议我停止不当用药，并明确声明自己不是医生，不能替代医生诊断（怕我sue它）；接着为我推荐了护理方式、饮食注意事项，并在我因为无法完成预期计划而焦躁时给予安慰。
Then I thought: why not ask AI? I sent it a photo and description. Its reaction was straightforward: firmly advised me to stop improper treatment, clearly stated it was not a doctor (probably to avoid lawsuits), then suggested care methods, diet adjustments, and even offered comfort when I felt anxious about disrupted plans.
炎症很快被控制住，随后出现好转迹象，完全恢复大概用了近两个月。我之前从没想过，AI 居然会在个人生活里帮我处理这种细碎的小事。而且，它还让我补了一些医学和药品的常识。而在跟它的对话与反思中，我也意识到：身体已经到了必须要有意识照料的阶段，个人护理也并非只是为了外表，而是为了健康的延续。
The inflammation was soon under control, then improved, and took about two months to fully heal. I had never thought AI could help with such trivial personal matters. It also taught me some medical basics. And through reflecting with it, I realized my body now needs conscious care — not just for appearance, but for sustaining health.
AI 虽然没有感受，但在信息检索、提醒错误操作与提供心理支持上，确实可以成为一种实用的陪伴。当然，责任依然在人类。
AI has no feelings, but in retrieving information, correcting mistakes, and offering support, it can indeed serve as a practical companion. Still, the responsibility remains with humans.
后来，它也逐渐进入到我的工作里。因为那块皮疹打乱了日常节奏，我第一次尝试让 ChatGPT 帮我做时间管理。结果发现，只要我给出清晰的要求，它就能把任务排布得很合理。我虽然很满意，但也有种怎么好像在给AI打工的感觉。
Later, it entered my work. Because that rash disrupted my routine, I first asked ChatGPT to help with time management. I found that with clear instructions, it arranged tasks well. I was satisfied, though I felt oddly like I was working for the AI.
然后，我让它参与到创作计划和研究主题的讨论里。它可以帮我进一步理清思路，也可以提供比较全面的背景信息，还能根据要求，给出一些不同的角度。
Then I brought it into discussions of creative plans and research topics. It helped me clarify ideas, offered background, and gave alternative perspectives when asked.
而AI作为一台算法机器，它最鲜明的特色是结构化，把模糊的想法变得清晰，把零散的念头汇聚到一起。这样固然有好处，但对创作而言，却也可能会成为问题。AI 的逻辑是条理化、可预测的，而创作却可能需要混乱、偶然和不确定性。如果过度依赖它，把一切都安排得清清楚楚，可能反而会削弱那些意外的火花。因此，人类的主体性更显得重要。
As a machine, AI’s strongest trait is structure: turning vague ideas into clarity, gathering scattered thoughts. That’s useful, but in creativity it can be a problem. AI’s logic is ordered and predictable, while creativity often thrives on chaos, accident, and uncertainty. If you depend too much on it, you may lose those sparks. Hence, human subjectivity is crucial.
除了工作相关内容，我也会与AI讨论一些社会事件。我会将我的思考过程与我的观点发给它，而它通常表示赞同，并为我的观点提供论据或补充。
Besides work, I also discuss social issues with AI. I share my thoughts and views, and it usually agrees, adding supporting arguments.
这里就浮现出一个关键问题，如果说在与AI讨论自己创作与研究时，它会帮你捕捉并锚定你作为个体的独特想法，而在与其讨论社会事件时，它又总会支持你的观点，那么AI的使用者是不是很有可能会封闭在自己的世界里。
Here arises a key question: in creative discussions, AI can anchor your unique perspective; but in social debates, if it always supports you, might you close yourself off from the world?
如果一个人只想被认同，而不想被挑战，在工作或创作上，可能会陷入自我的困境，而在社会议题上，因其带来的回音室效应也尤其危险。
If someone only seeks affirmation and avoids challenge, they may get stuck in self-made traps. In social matters, this echo chamber effect is especially dangerous.
这种风险是真实存在的。我也看到过报道：有人在与生成式预训练语言模型互动时陷入错误的信念或依赖，或者执迷于自己的妄想，带来严重后果。
This risk is real. I’ve seen reports of people engaging with language models, falling into false beliefs or delusions, with serious consequences.
关于AI的角色，诺贝尔物理奖得主杰弗里·欣顿提出AI应该被赋予母性本能，而AI领域的学者李飞飞则认为应该把AI当作一个工具，究竟AI应该是母亲还是工具，我想这里有一个责任的问题。
On AI’s role: Nobel laureate Geoffrey Hinton suggested AI should be given maternal instincts, while Fei-Fei Li argued it should be seen as a tool. Mother or tool — this is really a question of responsibility.
物理学家薛定谔曾在一篇探讨意识的文章里提到过人类照顾自己的责任，如果人类不能照顾自己，那也很难在这一条通路上继续进化。我倾向于认为，AI是一个智慧体，而责任在人类自身。对个人来说，我们要决定如何使用 AI；对社会来说，我们要决定如何监督与规范它。
Physicist Erwin Schrödinger, in an essay on consciousness, wrote about the responsibility humans have to care for themselves — otherwise, further evolution may be impossible. I lean toward seeing AI as an intelligent entity, but responsibility lies with humans. As individuals, we must decide how to use AI; as a society, we must decide how to regulate it.
所以在使用 AI 时，要求人类必须有较强的判断力。AI 默认对使用者是支持性的；明确意识到这一点，我们更应该提醒自己在多个角度上展开思考；而在与AI的互动中，我们也应要求它提供多个角度，比如，我们可以用这样的提示词，请给出反方观点、列举证据、比较不同立场。
Thus, using AI demands human judgment. By default, AI supports the user; knowing this, we must remind ourselves to seek multiple angles. In interacting with AI, we can ask it for opposing views, evidence, or comparative perspectives.
通过有意识地让 AI 提供多重视角，我们反而可以用它来锻炼批判性思维，而不是让它变成遮蔽视野的镜子。以这样一种方式，我们可以始终保持思考的开放性，训练判断力的敏锐与流动性。
By consciously prompting AI to offer multiple perspectives, we can use it to train critical thinking, rather than let it become a mirror that narrows vision. This way, our thinking remains open, and our judgment sharp and flexible.
最后，再说回到艺术与创造力。有人担忧 AI 会取代艺术家，在我看来不会；如同 AI 不能取代人类的判断力，我认为 AI 也不会替代人类的创造力。使用得当，它反而可能会促进个体的创造力。
Lastly, back to art and creativity. Some fear AI will replace artists; I don’t think so. Just as it can’t replace human judgment, it won’t replace human creativity. Used well, it may even boost it.
确实 AI 能很好做规律性强的设计任务，但艺术的价值往往来自那些仍然没能找到规律的东西，比如，就绘画而言，个体的感受与经验所带来的潜意识的涌动，身体的手势，媒介材料的偶然性。
In my own experience, AI’s biggest help in creation is showing me my own ideas. When I input fragments, it organizes them, making me realize my thoughts were already quite complete. Or it throws out an unexpected idea, helping me break habits. But the real choices must still be mine.
AI 有各种用途，但真正的作品仍然需要人去做出那一笔、一处肌理、一种选择。
AI can inspire, provide variations, or serve as a creative catalyst. But the real work still requires humans to make that stroke, that texture, that choice.
谢谢你的收听，再见。Kwaheri [kwa-HE-ree]! (斯瓦西里语 再见) Miao~
Thanks for listening. see you next time. Kwaheri [kwa-HE-ree]! Miao~