Luyện nói tiếng Anh bằng Shadowing qua video: Could AI chatbots replace human therapists? - What in the World podcast, BBC World Service

I think that some therapists could see, you know, that AI could be a helpful sort of assistant.

⏸ Tạm dừng

Tốc độ:

146 câu

Nếu các câu quá ngắn hoặc quá dài, hãy bấm Edit để chỉnh sửa.

I think that some therapists could see, you know, that AI could be a helpful sort of assistant.

But I think there's a lot of therapists that are afraid.

Do you use an AI chatbot to talk about personal stuff?

I know that I have been relying on mine quite a lot recently.

Like, I can ask it, what can I do to reduce my stress?

And it says, here are some quick stress relief techniques: Deep breathing.

Muscle relaxation. Mindfulness. Take a walk.

And millions of people on TikTok are talking about using things like ChatGPT or DeepSeek as therapists.

There are, of course, specific therapy chatbots too, but there are some concerns about what happens to your data.

Is AI therapy safe and does it work?

And today we're going to go through the different therapy chatbots that are around.

And we're going to talk about whether they could be a replacement for traditional therapists.

I'm Hannah Gelbart and this is What in the World from the BBC World Service.

And here with me in the studio to talk about this is Jordan Dunbar. Hi Hannah.

Hi. So you have made loads of documentaries for the BBC about mental health.

What kind of AI therapy apps are out there, and have you used any yourself? So most of my documentaries have been looking at how you can get safe human mental health help looking at the therapy industry both in the UK and around the world.

But where it's sort of started to cross over is the fact that there's a major problem in many, many countries about accessing mental health care.

So psychiatrists and therapists are both really expensive to train.

There's not enough of them.

And it takes years when you get people who do want to sign up.

So it means there's a real shortfall.

Plus, we're all talking about our mental health a lot more, and that's been really good because it's meant that people have come forward and it's sort of getting rid of the stigma in a lot of countries about talking about it. However, the trade off is that means more people need psychiatrists and therapists and they're not there.

And if they are there privately, they can be really expensive.

So people in my documentaries have then been looking at online therapy, like AI that we've been talking about. Now, generic AI, so the one that you might go, like a search chat bot, that you would get help for booking tickets as well as using as a therapist, or dedicated therapy apps.

So those are the ones built from the ground up to help with your mental health.

So the one that I tried and this a few years ago was quite a basic one.

It was more like journaling. So it would ask about your thoughts.

How's your day been? How are you feeling about that?

And the idea was, a bit like how you would never, I think, speak to a friend, is getting you to think about what's happened in the same way that you might keep a diary.

And that's quite helpful because you can scroll back and remember, oh, I felt like this the other day, or that really annoyed me.

And it helps you sort of process your thoughts, but I never used it for anything really serious.

It was more just like, yeah, like more of an online diary, looking at low mood.

As you say, lots of people are using them because there's no waiting lists.

They're cheap or free.

And of course they're like in your pocket if you need them.

And we have been hearing from people who have used them.

This is Michael from Austria.

I've been having some anxiety, like money anxiety, and it's been ruminating in my head like all the time.

I wanted a solution and I went to ChatGPT because I was using that for work anyway.

And I was like, just asking, hey, is this normal?

What's wrong? Why do I have keep having those thoughts?

And he gave me like a real good plan, like a seven-day plan, um, that I followed, like things, exercises, uh, things that I need to think about.

And it actually worked. When I go to a therapist, it costs me 100 euros per hour. Maybe it helps. Maybe it does not.

It's quite some investment here, but that's not the problem.

I need to wait four to six weeks.

Because in Austria, it's really hard to find a therapy place.

With AI, I'll just go in there and discuss it. I might get a solution, I might not, but usually after that, I'm feeling better than I did before. Because this, uh, those things really show empathy.

Like, it's, uh, really getting close, like conversing with a human.

And I hate to say it, but maybe sometimes it's more empathetic than a therapist.

So, Jordan, I've used, like, a generic chatbot as a therapist, and it's quite good for the basic stuff, although a little bit cheerleadery and very positive.

Too positive for you. You want a more miserable chatbot! Be more realistic!

I want realistic, I want realism.

Yeah, I want someone who listens and reflects and not just says, like, you're doing fine, you're doing great.

Um, and I'm wondering if there might be a difference here between the generic chatbots, which are like personal cheerleaders in your pockets and the more dedicated therapy apps.

How are they built and how do they differ?

Well, a lot of the big ones certainly claim that they've been built with mental health professionals, and a lot of them use a type of therapy called CBT that's cognitive behavioural therapy.

So whereas CBT is slightly different from what you might have seen on TV or talking therapies, it is a lot more structured.

So when I've done it, it helps you try and reframe your thoughts, particularly if you're dealing with anxiety or low mood.

So it's trying to get you to think differently so you don't get caught up in those spirals. So it's very structured, which means that works quite well with an online app.

It can ask you certain questions, it can prompt you, but it's not really about giving advice.

It's more about getting you to think about your own thought process.

Some are based on large language models, so that's where it's taking loads of different data from the past and from the internet. Tried to learn from that.

And then whatever you write in based on the previous data, it tries to give you answers or further prompts.

The only problem with that is, most therapy that we've had for hundreds of years, that's all been written down by psychiatrists and therapists.

They're not sharing that, right. That's not online on the internet. - That's personal data isn't it? - Exactly.

For models to learn from. So we don't actually know exactly where these large language models are learning.

And obviously every app is different.

And there have been some studies into how effective these therapy chatbots are.

Yeah. Well it is very, very early days and there are many, many, many different apps, which means it's difficult to get reliable evidence.

There is one from the New England Journal of Medicine, and it said the first clinical trial of a therapy bot that uses generative AI suggests it was as effective as human therapy for participants with depression, anxiety or risk for developing eating disorders.

But the model was trained on the scientists' own custom data sets that was based on evidence based practice and not just general mental health conversations, and had a team of trained researchers to monitor interactions.

But as it said, it's on their custom data sets.

Essentially, what we really need is to be able to have a long enough time with people from all over the world and all walks of life, really diverse sets of people to understand what this works for.

When we say mental health, we mean from low mood, which is completely normal, right through to really serious illnesses like schizophrenia.

So we talk about mental health, but it's a huge spectrum.

And going to different apps might be beneficial for some things.

Low mood, anxiety, being able to talk to someone, particularly if you're in a culture where you couldn't open up to a friend or you can't afford therapy, but that's very different from something like psychosis or trauma or abuse.

So this evidence says that it's for things like depression and anxiety, but lots of it will be things that are more treatable through speaking rather than things that you need medicine for and really, really trained professionals.

The cultural experience of therapy is so important in this because different cultures have different ways of accessing therapy.

There's different stigma about speaking about your problems and your personal issues.

And I was really interested in the rise of the Chinese AI app DeepSeek, because more young people in China have been using it for emotional support.

And we heard from our reporter Kelly Ng, about this.

There's this lady in her 20s who told me that she would use DeepSeek for nightly therapy sessions before she goes to sleep.

Um, and she talked about how it had comforted and encouraged her in a way, the kind of encouragement that it gives her is very different from her friends and family.

And before the app sends its reply and its suggestions and advice, it will first show you its thought processes.

So this can go from like it analysing your query and and you know, talking about what you are actually really trying to ask.

And then it also gives itself reminders like to be empathetic, to reply in a certain tone, use certain words, and so on.

Um, and I think that has struck a chord with some of the users who feel like they are talking to a fellow human, like a, like a friend, a friend who, um, remembers everything that they say because it's a robot after all.

I found statistics from 2022 where out of a million Chinese people, only 20 have access to mental health services.

So you can imagine what a small proportion that is.

And then, apart from physical and financial access, there are also cultural barriers.

So, for instance, conversations about mental health remain a taboo in many Chinese families or indeed many Asian families.

And that's why they may prefer using apps like DeepSeek because of that sense of privacy and anonymity.

So what do you professional therapists and psychologists make of these?

So I think lots of them admit that there's a problem in terms of supply and cost.

In places across the world. For human therapy.

For human therapy, and therefore having something that's really easy to access, that particularly young people who are digital natives and are happy to go online and also maybe feel more comfortable sharing their feelings with rather than a stranger.

It's so uncomfortable, isn't it?

Having therapy sitting in that room or on zoom?

100

Yeah, well, it's like dating, isn't it?

101

You've got to find the right person you have to go through.

102

It's also quite complicated.

103

So to have something on your phone that's cheaper and easier.

104

Lots of them think that's a good thing if it helps break down the stigma.

105

I think where they think it's a bad thing is if you're trying to directly replace human therapy.

106

And we reached out to a psychologist in the US. This is Maytal Eyal.

107

I think that some therapists could see, you know, that AI could be a helpful sort of assistant that can be there for clients outside of therapy sessions.

108

But I think there's a lot of therapists that are afraid.

109

I think as people become increasingly more reliant on a chatbot for emotional support, it's going to create a different expectation for intimacy and for connection and for support.

110

That becomes an impossible standard for other humans to live up to.

111

Endless patience and constant availability and total selflessness of a bot is unachievable in our other human relationships.

112

If we get used to being in relationship with an AI bot, whether it's like our therapist that we talk to 24/7, or a friend, we kind of will forget that natural friction that's inherent to intimacy.

113

We'll forget how to do that. Our capacity to tolerate discomfort might atrophy away, and that's really scary.

114

But Jordan, surely any of these AI bots, whether they are the dedicated ones or the more kind of generic ones, could get it wrong and give harmful advice?

115

Yeah, and I think you have to remember that therapy and psychiatry has developed over such a long time.

116

And humans are incredibly, incredibly complicated, right?

117

We work on so many different signs and nuances.

118

Those aren't going to get picked up by an app.

119

It'll take a very long time to be able to train them to get to that point.

120

So if you're dealing with something like self-harm, suicidal thoughts, drug abuse or particularly trauma, those are things that really have to be dealt with by a professional.

121

Have you come across any horror stories?

122

Well, in 2023, the National Eating Disorders Association US, it replaced its live helpline with humans.

123

It replaced it with a chatbot. And then it had to suspend the chatbot because people using it were claiming that the bot was recommending calorie restriction, which is not correct for eating disorders at all.

124

So as you can imagine, if these are going to replace humans in lots of different times, they are going to get things wrong.

125

I think the most important thing about this is, what happens when that goes wrong?

126

So if you were with a therapist or a psychiatrist and something goes wrong, you want to make a complaint. In most countries, the regulations where there's a body, there's a place you can go to complain.

127

But on an app, what happens?

128

That's another thing you should really be looking out for is what does it say if you need more help or things go wrong?

129

So Jordan, if people are using gen AI to talk about their feelings and this deeply personal stuff, mental health, emotional issues, is there a risk that that data, that information about them, could be leaked?

130

I mean, it could be really valuable for marketers who want to, you know, sell you stuff to make you feel better if you've got low mood.

131

Yeah, it's quite scary thinking if you're sharing your innermost thoughts, particularly if you're dealing with things around anxiety and fears that you're putting that out there, you really want to know how that data is being used.

132

If you're using a generic AI, like one of the search chatbots, that's much more difficult because lots of the data that's put in the questions that people ask that's used to train the models.

133

If you're using a mental health app that has been designed specifically for mental health.

134

It should say on it in the terms and conditions what the privacy is.

135

Is it going to be sold to third parties, who can access it, are their mental health professionals involved?

136

And actually, the Mozilla Foundation.

137

So they're a global nonprofit, and they attempt to police the internet looking for bad actors, as they call them.

138

So they surveyed 32 leading mental health apps for a report last year.

139

They found 19 of them were failing to protect user privacy and security.

140

So it is really worthwhile. I know it's boring, but reading the small print to work out how your data is going to be used and what your privacy is.

141

So just to summarise, the experts that we have been speaking to for this have said that I shouldn't be used as a replacement for therapy.

142

Some of the AI companies have said that users should also seek professional help, because these bots are, of course, not licensed therapists. Jordan, thank you so much for joining us.

143

Thank you. And that is it for today's episode.

144

Thank you for joining us. I'm Hannah Gelbart.

145

This is What in the World from the BBC World Service.

146

We'll see you next time.

TRENDING

Phổ biến

Phương Pháp Shadowing Là Gì?

Shadowing là kỹ thuật học ngôn ngữ có cơ sở khoa học, ban đầu được phát triển cho chương trình đào tạo phiên dịch viên chuyên nghiệp và được phổ biến rộng rãi bởi nhà đa ngôn ngữ học Dr. Alexander Arguelles. Nguyên lý cốt lõi đơn giản nhưng cực kỳ hiệu quả: bạn nghe tiếng Anh của người bản xứ và lặp lại to ngay lập tức — như một "cái bóng" (shadow) đuổi theo người nói với độ trễ chỉ 1–2 giây. Khác với luyện ngữ pháp hay học từ vựng bị động, Shadowing buộc não bộ và cơ miệng phải đồng thời xử lý và tái tạo ngôn ngữ thực tế. Các nghiên cứu khoa học xác nhận phương pháp này cải thiện đáng kể phát âm, ngữ điệu, nhịp điệu, nối âm, kỹ năng nghe và độ lưu loát khi nói — đặc biệt hiệu quả cho người luyện IELTS Speaking và muốn giao tiếp tiếng Anh tự nhiên như người bản ngữ.

ShadowingEnglish.com – Luyện Nói Tiếng Anh Bằng Shadowing

Nói tiếng Anh lưu loát với phương pháp Shadowing — nghe người bản xứ, đọc theo từng câu, rèn phản xạ ngôn ngữ thật sự. Được hàng ngàn người học IELTS tin dùng.

Luyện nói tiếng Anh bằng Shadowing qua video: Could AI chatbots replace human therapists? - What in the World podcast, BBC World Service

Phổ biến

Tại sao nên luyện nói với video này?

Ngữ pháp & Cách diễn đạt trong bối cảnh

Các cạm bẫy phát âm phổ biến

Phương Pháp Shadowing Là Gì?

Luyện nói tiếng Anh bằng Shadowing qua video: Could AI chatbots replace human therapists? - What in the World podcast, BBC World Service

Tải Ứng Dụng

Tại sao nên luyện nói với video này?

Ngữ pháp & Cách diễn đạt trong bối cảnh

Các cạm bẫy phát âm phổ biến

Phương Pháp Shadowing Là Gì?