Name: 麥可波倫：我不認為 AI 會產生意識！ (I Don't Think That AI Will Become Conscious - Michael Pollan)
Uploaded: 2026-02-20T17:19:53.000Z
Duration: 5 min 34 s

We're the author of the new book A World Appears.

我們是新書《一個世界出現了》的作者。

A lot of people are worried about AI and consciousness.

And do you share those worries or do you have share.

你也有這些擔心嗎？還是你有自己的擔心？

Yeah, there's enough things to be worried about.

So I don't think that AIs will become conscious.

The belief that they can become conscious is based on a faulty metaphor.

認為它們能產生意識的信念，是基於一個錯誤的比喻。

The metaphor that brains are computers and brains are not like computers at all.

大腦是電腦的比喻，但大腦根本不像電腦。

You, my friend, have never read Dianetics.

我的朋友，你從來沒讀過達觀學。

Brains, you know, in computers, you have this separation between hardware and software, right?

大腦，你知道，電腦裡，你會區分 hardware 和 software，對吧？

And you can run the same software on any number of different computers.

而且你可以在任何一臺不同的電腦上執行相同的 software。

Every memory, every experience you had physically rewires your brain, changes the matter of your brain.

你的每一個記憶、每一個經歷，都會實際地重塑你的大腦，改變你大腦的物質。

So our brains are not interchangeable because we had different life experiences that shaped them.

所以我們的大腦是無法互相替換的，因為我們有不同的生命經歷塑造了它們。

So the experience is not necessarily just stored there, but

所以經歷不一定只是儲存在那裡，而是

it actually changes the environment, the matter which is stored.

它實際上改變了儲存它的環境、物質。

When you're young, you have many more connections, and they gradually get pruned.

你年輕的時候，神經連接比較多，然後會慢慢被修剪掉。

And the way they're pruned depends on your life experience.

而它們被修剪的方式，就取決於你的人生經驗。

The other problem is that it appears that consciousness begins not with thought, higher order thought in the cortex, but

另一個問題是，看起來意識的產生，不是來自大腦皮質的高階思考，而是

它們現在會思考，但卻無法感受。

Since feeling is first, whoever pays attention to the syntax of things will never wholly kiss you wholly.

因為感覺至上，任何關注事物語法的人，都無法全然地親吻你。

To be a fool While spring is in the world.

在春天還在世界上的時候，當個傻瓜。

My blood approves, My lady, I swear by all roses, the best gesture of my brain is less than your eyelids flutter.

我的血液贊同，我的女士，我對所有玫瑰發誓，我腦袋裡最好的舉動，都不及你眼皮的輕顫。

那說明瞭我們是為彼此而存在的。

But I actually do believe in almost all human behavior.

但我其實相信幾乎所有的人類行為。

How do you differentiate between feelings and thoughts?

Because I always thought that feelings are just sort of the unseen and vast connections between so many thoughts and memories and experiences and our anticipations of the world that meet like the mycelium under the forest floor where we can't see them and exchange information between these thoughts, ideas and memories and produce this sort of aura that we call a feeling.

因為我一直覺得，感覺就像是很多想法、記憶、經驗和我們對世界的預期之間，那些看不見的、廣闊的連結，它們在地底下像菌絲一樣交織，我們看不見它們，但它們在這些想法、點子和記憶之間交換資訊，然後產生一種我們稱為「感覺」的光環。

I guess I say that to ask the question, what is a feeling?

我這麼說，其實是想問，什麼是感覺？

Feelings are how the body talks to the brain.

感覺是身體跟大腦溝通的方式。

We forget brains exist to keep bodies alive, not the other way around.

So the brain is tracking what's going on.

How's the temperature, your blood gases, all this kind of stuff.

And when things go off and you're out of homeostasis, a signal comes that is felt by the brain as a problem or something.

我們常常忘了大腦其實是為了讓身體活著，而不是反過來。

So I think you need a body to have feelings.

所以大腦就是在追蹤身體裡發生的一切。

I don't think you can do feelings without a body.

I think you can't do feelings, feelings without having a vulnerability, probably without being mortal.

像是溫度、你的血液氣體，這類所有東西的狀況。

I mean, the fact that we can suffer, our feelings are rooted in this bodily existence.

當身體失衡，偏離了恆定狀態，就會發出一個訊號，大腦會感覺到這是一個問題，或是有什麼狀況。

而且我不知道機器要怎麼才能達到那樣的境界。

Have you spent much time speaking with digital Personas like Claude or Gemini or ChatGPT?

你有花很多時間跟 Claude、Gemini 或 ChatGPT 這類的數位 Persona 哈拉嗎？

Because I literally have not typed one thing into any of them.

因為我真的什麼都沒打過任何一個字進去。

They, you know, first of all, they talk to us in the first person.

牠們，你知道的，首先，牠們會用第一人稱跟我們說話。

And, you know, they're good for some things.

而且，你知道的，牠們在某些方面做得很好。

I've used them in research, and sometimes the results are just so wrong, it's like, laughable.

我曾用它們做研究，有時候結果錯得太離譜，簡直讓人笑掉大牙。

But I really think that these just chatbots, you know, 72% of teenagers in America now turn to AI for companionship.

但我真的認為這些只是哈拉機器人，你知道的，現在美國有 72% 的青少年會尋求 AI 的陪伴。

Even though chatbots are not going to become conscious, they're going to fool a lot of people, and

Well, Michael, before you go, I just want to say thank you for being here.

雖然哈拉機器人不會產生意識，但它們會欺騙很多人，

You were one of my favorite guests on the old show.

You're still one of my favorite guests here.

麥可，在你離開之前，我想謝謝你來這裡。

Whenever I would see your name on a card up on the board of the people I got to talk to that week, I always got happy and excited that I knew we'd have a great conversation.

I just want to thank you for all those conversations the last 20 years.