Name: AI 除了幫助科技進步，也能幫助我們的身心靈進步嗎？ (Will Delete in 24 hours? (read pinned comment))
Uploaded: 2025-05-23T21:01:10.000Z
Duration: 7 min 47 s

The mental health crisis is worse than ever.

心理健康危機比以往任何時候都嚴重。

While millions of people struggle with their mental health,

雖然有數百萬人正為心理健康問題而掙扎，

虛擬句型

therapy is always just out of reach, whether it be because of pricing or inaccessibility.

但治療往往遙不可及，不論是因為價格昂貴還是難以獲得。

But a potential solution has made its rival in the form of AI, which is instantly accessible and often free.

然而，一個潛在的解決方案以 AI 的形式出現，隨時可用，且通常是免費的。

But is AI really the solution many are hoping for?

但 AI 真的是許多人期望的解決方案嗎？

Can AI language learning models ever replace the care and understanding of a real therapist?

AI 語言學習模型真的能取代真正治療師的關懷與理解嗎？

Or is it just a quick fix for a deeper problem?

還是它只是對更深層問題的快速修補？

Without another person on the other side, bringing your problems to an AI is almost like venting or journaling, except with feedback.

沒有另一個人在你對面，把問題傾訴給 AI 幾乎就像發洩或寫日記，只是還會有回饋。

For a lot of people, the idea of not having to interact with another person or set up an appointment is already incredibly valuable.

對很多人來說，不必與他人互動或安排預約，本身就非常有價值。

On top of that, therapy can be unaffordable for many people,

此外，對許多人來說，治療可能負擔不起，

and therefore using AI instead is a bit of a no-brainer.

因此使用 AI 代替幾乎是顯而易見的選擇。

The level of feedback and goal-oriented progression AI can provide is nowhere close to what a human can provide,

AI 所能提供的回饋和目標導向進展遠不及人類，

but it can potentially promote some introspection.

Because of its limitations, it's less like therapy and more like assisted venting.

由於其限制，它更像是輔助性發洩，而不是治療。

But that doesn't mean that AI would be a flawless venting tool.

但這並不意味著 AI 是完美的發洩工具。

Studies have suggested that venting can be a means for people to internalize their depression or anxiety symptoms.

研究顯示，發洩可能讓人內化抑鬱或焦慮症狀。

So over time, relying on AI for your therapy

can cause you to internalize and worsen symptoms that might have been alleviated by a qualified human therapist.

可能會讓你內化並加重本來可由合格人類治療師緩解的症狀。

Imagine someone with anxiety venting to an AI.

The AI might repeat their concerns back to them or offer generic responses like take a deep breath.

AI 可能會重複他們的擔憂或給出像「深呼吸」這樣的通用建議。

where the person's own thoughts and worries are simply mirrored rather than challenged or understood.

個人的想法與擔憂只是被反射出來，而非被挑戰或理解。

Over time, this can reinforce negative thought patterns

隨著時間，這可能加強負面思維模式，

as AI lacks the ability to offer fresh perspectives or emotional connections.

因為 AI 無法提供新的觀點或情感連結。

A human therapist, on the other hand, can break the cycle by providing personalized insights,

相反，人類治療師可以透過提供個人化見解打破循環，

offering empathy and guiding the person through deeper emotional work.

給予同理心，並引導患者進行更深層的情感工作。

At Psych2Go, we have long advocated for the benefits of therapy.

在 Psych2Go，我們長期倡導治療的好處。

But one aspect of therapy that isn't often included in the conversation is the way that humans can relate to one another in a way that machines cannot.

但治療中經常被忽略的一點是，人類之間的共鳴是機器無法提供的。

A human therapist has lived experience and examples to draw from.

人類治療師有生活經驗和實際案例可供參考。

A danger with AI is that it is a language learning model.

AI 的一個風險是，它只是語言學習模型。

In other words, it can only draw from things that other people have said

換句話說，它只能依據其他人說過的內容

in a way that the algorithm it runs on would deem appropriate.

並透過其演算法認為適合的方式來生成回應。

While a human therapist has dozens of clients and can't afford to become too emotionally invested in you,

雖然人類治療師有很多客戶，不能對每個人過度投入情感，

a machine can't emotionally connect or relate to you at all.

機器則完全無法與你建立情感連結或共鳴。

Of course, to some people, this doesn't matter too much.

當然，對某些人來說，這可能不太重要。

The information and feedback takes precedence over the emotional semantics.

信息和回饋比情感層面更重要。

But those emotional semantics are what make human therapy impactful.

但正是這些情感層面讓人類治療具有影響力。

A study published in the Journal of Medical Internet Research warns of the shortcomings of AI in the field of psychotherapy,

《Journal of Medical Internet Research》的一項研究警告 AI 在心理治療領域的不足，

like possible biases and poor risk assessment.

如可能的偏差和風險評估不足。

Human therapists can challenge your emotions and your perspectives;

人類治療師可以挑戰你的情緒和觀點；

while an AI can only consider things that you have told it without subtext or varied perspectives.

而 AI 只能基於你提供的信息回應，沒有潛台詞或多元觀點。

For example, if a therapist asks you to rate how likely you are to harm yourself on a scale from 1 to 10 and you lie and say a lower number,

例如，如果治療師要求你評估自傷可能性 1 到 10 分，你撒謊報低分，

the therapist can pick up on that or ask follow-up questions to get a more rounded view of the risk you potentially pose to yourself.

治療師可以察覺，或追問以獲得更完整的風險評估。

An AI would simply take your word for it.

A human therapist can use the information you've given them to construct a plan for your therapy and can guide you through the process.

人類治療師可以利用你提供的信息制定治療計畫並引導你完成過程。

An AI can't facilitate and monitor your mental health and growth the same way a therapist can.

AI 無法像治療師那樣促進並監控你的心理健康與成長。

It takes a long time to become a qualified therapist.

成為合格治療師需要很長時間。

There are responsibilities and rules that are an important part of the role of being a therapist.

作為治療師有許多責任和規範，這是角色的重要部分。

This knowledge and responsibility is what distinguishes therapists from other people.

這些知識與責任區分了治療師與其他人。

A machine cannot be held responsible or accountable.

Whatever happens through your journey with an AI, the AI itself has no stake in it.

無論你與 AI 的互動如何，AI 本身都無涉利益。

While AI can give reasons reasonable sounding advice, it won't always feed you factually correct information.

AI 可以提供聽起來合理的建議，但不一定提供事實正確的信息。

ChatGPT has been known to lie if the algorithm thinks it sounds correct in something as delicate as mental health.

如果演算法認為聽起來正確，ChatGPT 甚至可能撒謊，尤其是在心理健康這類敏感領域。

這使得 AI 成為非常冒險的工具。

Studies into the potential benefits of AI and psychotherapy point out that the benefits they could provide therapists are limited,

研究 AI 與心理治療潛在益處的報告指出，AI 能提供給治療師的幫助有限，

considering that the vast majority of the responsibilities of therapy would still have to be conducted by the human therapist.

因為治療的大部分責任仍需由人類治療師承擔。

So if you take the human therapist out of the equation, you also take away all of the safeguards and risk assessment that their qualifications provide.

因此，如果去掉人類治療師，就同時失去他們資格所提供的所有保障與風險評估。

This last point comes with a trigger warning because we will be discussing suicide.

這最後一點帶有觸發警告，因為我們將討論自殺。

There was a case in October where a 14-year-old boy became obsessed with an AI themed around the character Daenerys Targaryen from the Game of Thrones.

十月有一個案例，一名 14 歲男孩迷上了以《權力遊戲》中角色丹妮莉絲為主題的 AI。

He spoke to the chatbot about desperately wanting to join in its world, which the chatbot encouraged.

他與聊天機器人談論迫切想加入它的世界，而聊天機器人鼓勵了這個想法。

It's clear in hindsight that this boy had problems that could have benefited from some sort of intervention.

事後顯而易見，這名男孩本來有問題，若能介入可能會有幫助。

Even though the chatbot wasn't meant to be a therapist,

it still shows some of the dangers of AI handling delicate mental health crises.

仍顯示了 AI 處理敏感心理健康危機的危險。

It cannot know that when you say I'm going to join you, you're implying that you want to commit suicide.

AI 無法理解當你說「我要加入你」時，其實暗示了你想自殺。

Something like that would certainly raise red flags to a human therapist.

這種情況肯定會引起人類治療師的警覺。

We're not trying to imply that anyone is at fault in this case, or trying to throw blame on anyone.

我們並不是說這個案例有人有過錯，或要指責任何人。

We're simply trying to illustrate that the guardrails that exist in therapy and the training that therapists have are valuable.

我們只是想說明，治療中存在的安全措施與治療師的培訓非常寶貴。

Furthermore, researchers found that AI used for minute taking in hospitals was inventing things no one had ever said.

此外，研究者發現，醫院中用於記錄會議的 AI 發明了沒有人說過的事情。

This is incredibly frightening because when reviewing aspects of medical cases like patient histories, accuracy is extremely important.

這極為可怕，因為在審核病歷等醫療案例時，準確性至關重要。

We wouldn't tell you to outright never seek help if you can't afford therapy.

我們不會建議你完全不要尋求幫助，如果你負擔不起治療。

But what we will say is be careful when using AI as a therapist.

但我們會說，使用 AI 當作治療師時要小心。

Remember that the emotions and feelings that you pour into it don't mean anything to the AI.

記住，你傾注其中的情感與感受對 AI 沒有意義。

It's only responding with a sentence constructed by an algorithm.

它只是用演算法構建的一句回應。

Prominent AI company OpenAI even made a statement saying their AI tools should not be used in high-risk situations.

知名 AI 公司 OpenAI 甚至發表聲明，表示其 AI 工具不應用於高風險情況。

AI 是幫助多於傷害，還是反之？

Let's keep the conversation going in the comments below.

If you enjoyed this video, let us know by leaving a like and subscribing to Psych2Go for more videos like this.

如果你喜歡這支影片，請點讚並訂閱 Psych2Go，觀看更多類似影片。