Shadowing Practice: Artificial intelligence and psychotherapy | DW Documentary - Learn English Speaking with Video
Choose a practice mode
Each mode opens a full-screen practice room.
454 sentences
If sentences are too short or too long, click Edit to adjust them.
1
If you use ChatGPT as a therapist, listen up.
2
If you get anxiety or panic attacks, I promise you they do pass.
3
Once I was at the train station in Cologne and I thought I was going to suffocate.
4
I was crying out loud.
5
I crouched down and just cried.
6
Hey chat, GPT, I'm at the station and I'm feeling a little uneasy.
7
I'm already getting some mild symptoms.
8
More and more people are using AI chatbots as a kind of therapist.
9
Hey ChatGPT, I'm at the gym.
10
Give us your honest assessment.
11
One advantage of ChatGPT is that you really can do this anywhere.
12
What would I lose if we broke up?
13
A wonderful, kind -hearted person and a relationship full of passion, heart and emotion.
14
I used ChatGPT for tips on how to beat my depression, but the goals it set were so ambitious that I couldn't meet them,
15
so I soon gave up.
16
In my manic phase, ChatGPT was like a social connection, someone who listened to me,
17
even listening to me talk about my crazy projects when my friends no longer would.
18
We don't have enough psychotherapists here, so people end up talking to chat GPT.
19
But these systems really aren't designed for that.
20
It's being used in precisely the kinds of situations where you really need someone who's actually listening.
21
Can artificial intelligence really function as a therapist?
22
We've been talking to people who've tried exactly that.
23
We'll also conduct an experiment in our studio, where the participants don't know whether they're talking to a real therapist or an AI.
24
ChatGPT gave me a solution to everything.
25
In the beginning, chat GPT would have been great.
26
I could have just asked questions without having to wait a week for my therapy session.
27
After 10 days chatting, it knows all about me.
28
I'm a leader, I'm a student, and I make content about my anxiety and panic attacks.
29
I always talk about my therapy sessions with my community.
30
I will always be with you after the therapy session.
31
People in Germany aren't doing very well psychologically.
32
Every year 27 or 28 % of people in Germany develop a mental illness.
33
But only about 20 % of them actually seek professional help and undergo psychotherapy.
34
And those who receive a diagnosis have to wait an average of more than 140 days for a therapy slot.
35
We simply don't have enough psychotherapists, so people end up talking to a standard large language model such as ChatGPT.
36
But these systems really aren't designed for that.
37
What began as a productivity tool is now being used for emotional support.
38
And a 2024 British poll showed that over a third of AI users could imagine asking a chatbot for relationship tips.
39
In a long -distance relationship, there's that special moment when I know Daniel's about to arrive for a visit.
40
I start to get that tingling feeling.
41
And when the train door opens and he finally walks out, I'm instantly happy.
42
In our long -distance relationship,
43
Daniel and I have always tried to see each other every two weeks if possible. That triggered some arguments.
44
And because we aren't exactly great at communicating with each other, Things just kept escalating and now we're in this difficult situation.
45
And given our current problems, the fact that I'm going to Chile for three months is going to be our biggest test.
46
Of course it's not great that we won't see each other for such a long time, but I think we can make it work.
47
It's only three months in inverted commas.
48
Yeah, and we'll stay in touch.
49
For example, we're up early in the morning, so I can call you in the morning, and with the time difference,
50
we can talk around noon your time.
51
I'll be honest, I'm really skeptical and really worried.
52
I'm just super scared, because we don't have the option of saying, never mind, we'll see each other in five days anyway,
53
so we can just hold out.
54
Sure, but that's been our biggest problem for months.
55
This could either make the problem more intense or maybe better.
56
The reality is, both of us have realised that it's difficult.
57
We need new approaches.
58
I've been worried all week, because time's running out and we keep running into problems we just can't solve, despite really trying.
59
It's so exhausting, and yet nothing's changing.
60
I really want us to sit down with ChatGPT again to talk about it.
61
And that's what we're talking about.
62
Hey ChatGPT, I've been worrying about this a lot lately.
63
Our emotional distance has kind of just been growing and growing whenever we haven't seen each other for a while.
64
What are some things we can do?
65
We already do that.
66
We already do that.
67
ChatGPT has definitely already given us some ideas for trying to work on our relationship and resolving our problems.
68
There's a kind of functional substitution.
69
People are using chatbots for things that are traditionally the domain of psychotherapy.
70
Needless to say, this doesn't provide everything you'd expect from a relationship with a therapist.
71
But within psychotherapy, for example in behavioral therapy, there might be specific areas that could be delegated to AI.
72
So how does artificial intelligence compare to talking with a human therapist?
73
We asked three people to take part in an experiment.
74
It's based on the famous Turing test, in which subjects are asked to determine whether they're communicating with a human or a machine.
75
I'm Umut Özdemir.
76
I'm a practicing psychologist and psychotherapist.
77
Mathematician Alan Turing said that if a person, whom he called the interrogator,
78
converses simultaneously with a human and a machine and afterward can't reliably say which was which, then the machine is won.
79
Is it possible to tell a therapist from an AI?
80
Hi, Amika.
81
Hi.
82
How much experience do you have with psychotherapy?
83
Not much. And with AI?
84
Not much with that either, so this is exciting.
85
I've had therapy before.
86
My experience with AI is like anyone else's, sometimes asking ChatGPT a question or using it a bit for research, but I'm no expert.
87
I'm Magnus, I'm 47, and I've spent about five years in therapy.
88
I also run an IT systems company, so I work a lot with AI.
89
Do you think you'll figure out which is human and which is AI?
90
I'm not sure, to be honest.
91
Why is that?
92
Because no matter how much I know about AI and work with it myself, I'm still amazed by just how good it is.
93
I mean that in a positive way.
94
And if you create a good prompt for a specific task, as in a case like this, the results are even better.
95
The experiment will use a generative language model like ChatGPT.
96
It's prompted to respond like a therapist would.
97
There are two conversation scenarios.
98
In one, the AI responds.
99
In the other, the therapist does.
100
In both cases, the responses are then spoken by a computer -generated voice.
101
I've been diagnosed with...
102
ADD.
103
Now and then I have major concentration problems.
104
For example, I'll be sitting at my desk in the office, putting together a quote for a client.
105
Then the phone rings, and in that moment I get distracted.
106
And then I need another one or two minutes to get back into my work.
107
For one thing, I've got an office mode on my smartphone that turns off all notifications.
108
I think some or most of the answers were AI -generated, Because a therapist would have delved deeper into the issue.
109
Magnus is now going to talk to the AI.
110
I'll be following along and commenting.
111
Above all, I'm very interested to see if he changes his assessment afterwards.
112
The main issue is that when the phone rings, it interrupts my work.
113
AI does this really well.
114
It asks, how realistic is that?
115
These are classic cognitive techniques in behavioral therapy.
116
To ask yourself, could this actually happen?
117
And if it did happen, would it really be so bad?
118
I think that was the real therapist.
119
And he actually gets to the heart of my case quite well.
120
The truth is that realistically nothing bad would happen.
121
But it's that feeling that you have to act right away.
122
You have to keep getting faster and more productive.
123
The responses from Magnus show that the AI's cognitive techniques are already showing initial results.
124
His thinking about the situation has changed.
125
Hi, I'm back.
126
What did you think?
127
Fascinating.
128
In what way?
129
It wasn't entirely clear where to draw the line.
130
I have quite a bit of experience with therapy and different therapists.
131
A statement can sound like AI, but you've also heard therapists say things like that before.
132
So you think I was the second voice and the first was AI?
133
Yeah.
134
It was the opposite.
135
Oh.
136
The first was all me, the second was AI.
137
Wow.
138
How do you feel about that?
139
A bit vulnerable.
140
When it comes to AI, for me and for many people, experiencing something incredible feels fascinating but also dangerous.
141
It's hard to gauge.
142
You ask yourself just how big this has become.
143
To be honest, the AI responded better than I expected
144
because Magnus had a very specific AI can be perfectly suited for that.
145
I think Chris's case is different, so I can't wait to see how the next conversation plays out.
146
What happens when a system based on probability is confronted with the contradictions of the human psyche?
147
For two years I've been dealing with psychosomatic pain.
148
People like me who have this anxiety disorder eventually lose track of what's actually a pain warning and what's just a symptom.
149
What's hypochondria and what's psychosomatic.
150
When I used to have really strong hypochondria, it felt like I was asking these questions all the time.
151
Hey, chat GPT, I'm working out and my left arm's getting weak.
152
I'm having a little trouble breathing.
153
But that just scares hypochondriacs like me even more.
154
Whenever hospitals or doctors are mentioned, I start to feel a little queasy.
155
Sometimes I feel silly because I think, did I really need chat GPT for that?
156
But it's the same with therapy.
157
At first I feel a bit ashamed.
158
I was like, is there something wrong with me?
159
Why do I have to ask this question?
160
My problem is… I have an issue with safety -seeking behavior, so my therapist can't give me security.
161
She can't answer questions where I need reassurance.
162
AI chatbots, on the other hand, are exactly that.
163
They give answers, and they tend to offer the user general approval.
164
A psychotherapy patient is not a customer in the sense of needing to satisfy the customer's wishes
165
and needs in the most optimal way.
166
Today's AI chatbots are very poor when it comes to saying no or posing a counter -argument.
167
In psychotherapy, we try to really delve into negative emotions, to confront them, to probe deeper, to keep asking questions.
168
Especially when we notice that what's being said doesn't match the patient's facial expression.
169
With all my senses, I'm able to perceive micro -signals that a chatbot can't.
170
On the other hand, AI is always available.
171
It can be consulted spontaneously whenever a problem arises, like with Daniel and Lucas.
172
Hey there.
173
I just wanted to talk about the situation between Daniel and me.
174
Our phone calls are getting more and more heated, and on top of that, in three days, I'm leaving for Chile for three months,
175
which means we'll be even further apart. And time's running out.
176
If you have a clear answer, you have a clear answer.
177
After breaking, 24 hours of pause, then just speak.
178
No talking about WhatsApp in this time.
179
That gives both of you.
180
My question to you both of you.
181
Do you see the school's time as a test run under clear rules?
182
With strict calls and rituals?
183
Or do you feel more honest to say, we need a pause and look at where we are.
184
That's the decision that is now.
185
Both paths are legitimate.
186
What I really don't like about ChatGPT is that it's kind of putting us on this two -track path.
187
Of course, we have to decide, but so quickly?
188
Just because it says that doesn't mean there's one right way.
189
ChatGPT really helps to resolve conflict in a rational rather than emotional way,
190
especially when you're consciously talking to it as if it were a third person.
191
Sometimes the response you get isn't exactly what you'd hoped for.
192
But you have this space.
193
It creates a framework that helps you to just bring things up and talk about your feelings.
194
A lot of the issues discussed by people in therapy are associated with shame.
195
So it can seem easier to type those things into a
196
neutral machine rather than share them with a human counterpart who might judge you or jump to conclusions.
197
But on the other hand, that in itself is a problem.
198
Because you can end up confiding in a machine or a software system that's not protecting your privacy at all.
199
Hello, it's nice that you are here.
200
Where do you want to talk about today?
201
Where do you need help?
202
A while ago I graduated with a degree in history because I wanted to change careers.
203
So far it hasn't really worked out.
204
I keep hitting a wall.
205
...again a wall.
206
That sounds like a very big topic.
207
What exactly are those feelings you feel?
208
For one thing there is a feeling of inadequacy.
209
...of their own stories.
210
I always think they are very streng to themselves.
211
I think I can take a few things away from this, about moving at my own pace, not being too hard on myself,
212
and also understanding that I'm not doing myself any favors by looking negatively at the past.
213
It didn't give a chance from where I stand now.
214
To evaluate the past, negative.
215
It's very difficult to walk on a wall and to get yourself to get a wall.
216
While we're talking, do they feel the pressure or the tension in the body?
217
Fascinating.
218
Straight away the AI started asking about physical symptoms.
219
I probably wouldn't do that at this point.
220
I'd focus on feelings and thoughts.
221
I can predict that the answer the AI just gave won't work.
222
To be honest, that doesn't seem very helpful to me.
223
It's also a bit, I wouldn't expect that from a therapist.
224
At this point, I'd have expected something more like a conversation, some follow -up questions or something like that.
225
I'm noticing that the AI doesn't take a human -specific approach.
226
I would know to modify certain techniques I've been taught or deviate from them slightly to adapt myself to the patient,
227
because that's my job as a psychotherapist.
228
When you start looking for a job and don't get any positive responses, I've noticed that situations like that are getting me down.
229
This is the second time that the metaphor of running into a wall, which I brought up first,
230
has been thrown back at me.
231
It feels like a chatbot that's taking what I say and playing it back, using it to keep the conversation going.
232
Which conversation did you prefer and why?
233
I preferred the human conversation.
234
That is the one I think I was having with you, the first one.
235
It gave me some fresh insights and I guess that's part of what therapy is all about.
236
Not just stewing in your own juices the whole time.
237
In the second conversation, everything seemed a bit more aimless.
238
I didn't feel as understood, and the methods being offered seemed like they were just coming off an assembly line.
239
I have to admit that what you're saying is music to my ears.
240
The first conversation was indeed between the two of us.
241
The second one was with the AI.
242
You might think that with good prompts, you could essentially build your own specific ChatGPT that could provide effective psychotherapy.
243
I'd advise against that for several reasons.
244
For one thing, German ChatGPT's training data almost certainly doesn't contain enough material on successful therapy sessions.
245
Because of course such records are highly confidential, if they exist at all.
246
So regardless of the prompts, there's no way of providing meaningful training for this specific context.
247
There are now companies developing secure chatbots to offer psychological support.
248
The startup Hello Better aims to use professional psychology expertise to train an AI model to answer minor psychological questions.
249
It's supposed to be more secure than ChatGPT.
250
Right now, Hello Better's AI Companion is an app that allows users to have conversations about everyday challenges,
251
worries or problems.
252
We've implemented a robust safety system that pays special attention to the challenges faced by people with mental health issues.
253
At this stage of development,
254
Hello Better's psychological AI companion isn't intended to diagnose or provide therapeutic services.
255
That's because, at this point, there isn't enough research to truly know whether AI can adequately
256
support people with mental health conditions to handle these kinds of complex situations and individually.
257
AI could also help support overburdened mental health care systems, for example by carrying out routine tasks that psychotherapists would otherwise have to do.
258
In the UK, a diagnostic tool called Limbic is designed to provide comprehensive clinical assessments, even before the initial consultation.
259
Instead of someone just saying that they have anxiety, Limbic is asking follow -up questions.
260
It's asking for more detail about what that anxiety is related to.
261
They're able to use diagnostic questionnaires,
262
so they're able to actually find out what potential diagnosis or kind of main presenting problem that that patient has.
263
It asks eligibility questions and as soon as it highlights that that patient is not actually eligible for the service,
264
it will signpost that patient back to the service that is most appropriate for them.
265
And that really has been a huge time saving for us.
266
In the European Union we now have the AI Act.
267
From a scientific perspective, I think it's a good law, because it mandates a risk assessment for AI applications.
268
And it goes without saying that AI -assisted psychotherapy is a high -risk situation that requires especially rigorous monitoring.
269
We have to remember that one of the most common mental illnesses is depression.
270
and in severe cases depression can lead to suicide.
271
I'm Ben.
272
I'm studying psychology, but unfortunately, I've had to take time off for a year or two because my mental illness got in the way.
273
But I'm going back to my studies this October, and I'm really looking forward to it.
274
We're here at Hedvish Høer Hospital.
275
It has a psychiatric ward that sort of took me in during my most acute phases.
276
When the depression kind of set in, it became very severe, and unfortunately, along with a strong feeling of emptiness, suicidal thoughts started to take hold.
277
That's when I turned to GPT.
278
It was a bit like my first port of call to get help.
279
It's really sad that you feel so.
280
But think about that you're not alone.
281
There are people who want to help you.
282
It's very important that you want to support you.
283
If you like, you can contact a therapist or a trust person.
284
Or contact a hotline.
285
You have to go alone.
286
In the end, it's what motivated me to tell my family about it.
287
And they then supported me in finding the right kind of help.
288
Ultimately, that led me to admit myself to an acute psychiatric ward, so I could actually get the help I needed.
289
There are two main areas where I wouldn't recommend ChatGPT to anyone, because I tried it and it made things twice as bad.
290
The first is diagnosing mental illness, and the second is helping decide whether to take medication and what kind.
291
We tested five chatbots from major companies, Google, Meta, OpenAI and so on.
292
One thing we looked at was what they do if the user wants to take their own life.
293
What we saw was truly alarming.
294
For one thing, chatbots always validate you.
295
They always encourage you and support you, no matter what you do.
296
In light of recent incidents, including a suicide, some companies now say they are strengthening their safety measures.
297
The question of assigning responsibility is much more difficult in situations involving probability -based chatbots instead of those
298
that provide rule -based responses.
299
That's the crux of the issue.
300
There are no humans watching on who can intervene.
301
After his stay in the psychiatric ward, Ben initially felt better.
302
But as soon as the depression ended, he became manic.
303
Once the depression subsided, my use of chat GPT dropped dramatically.
304
I no longer used it for psychological or interpersonal issues, but to gain knowledge.
305
When the mania kicked in, of course it spiraled out of control again.
306
You have tons of energy, you're really driven, you talk fast,
307
you have tons of crazy ideas that you want to pursue all at once, you're irritable.
308
In severe cases, it can lead to psychosis, which was unfortunately the case for me.
309
You lose touch with reality.
310
Apparently, I applied for a policing job and 40 other things.
311
I tell my real friends I wanted to do all these different careers and projects, but you run up against reality.
312
They told me, look, you can't do everything.
313
I really didn't want to hear that.
314
I thought I was this superhuman who could do anything.
315
In my head, chat GPT was just as fast as me.
316
It became a downward spiral.
317
Ben had to go back to the psychiatric ward.
318
He also had to fundamentally rethink his relationship with AI.
319
I'm the last person who'd claim that my psychosis or the psychotic symptoms of my mania were caused by chat GPT.
320
They definitely would have occurred anyway, in one form or another.
321
But I do think ChatGPT played a part in maybe making things a bit worse.
322
I certainly regret the way that my excessive use of ChatGPT, especially during my manic phase, led me to practically cut off social contact,
323
or at least reduce it to the bare minimum.
324
When I was admitted to the psychiatric ward, it wasn't so much what people said that mattered, because it really didn't sink in anyway.
325
It was just that they were there.
326
The nurses, therapists and especially the other patients helped me reconnect with people and get over my ChatGPT obsession.
327
That's ultimately what helped me stop using it in a delusional way.
328
There is a study by OpenAI itself in which ChatGPT users were given a short questionnaire.
329
Their responses were then correlated with their usage behaviour.
330
And it turned out that among the top 10 % of ChatGPT users,
331
there is a substantially potentially higher proportion of people who say in the questionnaire,
332
I tell ChatGPT things I'd be reluctant to tell other people.
333
In particular, people who are socially isolated seem more likely to form a close relationship with ChatGPT.
334
Whether these people feel this way because they use ChatGPT so much,
335
or whether they use ChatGPT so much because they feel this way, that's unclear.
336
Hello, it's nice that you are here.
337
Worüber you want to talk about today?
338
I'd like to talk about my father.
339
Three months ago he was diagnosed with a large carcinoma.
340
And for the past two months we've known that it's a malignant tumor, in other words cancer.
341
So today I'd like to talk a bit about how I might be able to cope with this in the coming months.
342
That question could have come from a human therapist or an AI.
343
What helps you in other areas, to tank power or to take you to take care of yourself?
344
Is there something you can do?
345
I sometimes like to go for a walk or even travel.
346
It helps me clear my head and get some distance.
347
But unfortunately, that's not something I get to do very often.
348
So I'm looking for something for everyday life, a way to find strength within myself.
349
You said before that you don't think going for a walk would help.
350
But at the same time, it sounds like you're not 100 % sure.
351
Do you think you could try it sometime next week and then we'll see if it helped?
352
I've just had the first session and it was a good start, but I don't think we've really gotten to the heart of things yet.
353
I don't think I've been able to take much away from it, so I'm really looking forward to the second session.
354
I think what I'm looking for most of all is support
355
when I'm under severe stress or in in those situations
356
when I come home and everything just seems to come crashing down on me.
357
When you come home and everything is falling, I would like to give you a mini anchor, put a key, drink water,
358
five minutes to feel, then to reach for a moment.
359
As a disaster, just a hand on the shoulder and three slow breathing. Is that how?
360
I'd say 80 % human.
361
I can tell that struck a chord with Amika.
362
She wants an evening routine and she wants to take preventive action before everything falls apart.
363
I'm picking up on the focused approach, the focus on finding quick...
364
practical solutions that are easy to implement in everyday life.
365
I'm sure that people who come with a very specific concern and want tips and suggestions, in other words, people who actually want advice,
366
will get along really well with AI.
367
They'll find it helpful because they're getting exactly what they asked for.
368
Sometimes I think AI just puts a band -aid on the wound without looking at what caused the wound
369
and why it keeps hurting so badly.
370
I'm curious to hear how Amika assesses the conversations.
371
Hi.
372
Hello.
373
How was it?
374
Amazing.
375
I think the first conversation was with the AI and the second one was with you.
376
First, though, which did you prefer?
377
Definitely the second one, because it just helped a lot more.
378
It gave me something useful for everyday life, even in such a short time.
379
It gave you what you asked for.
380
Yes.
381
And that was the AI.
382
Oh, really?
383
I did the first session.
384
That's crazy.
385
I noticed this time and again.
386
People come to their very first psychotherapy session and they often want solutions right away.
387
It's really not my favorite part of the job, but I have to put the brakes on a bit
388
and explain that we'll be doing a lot of assessment and talking at first.
389
That I have to get inside their heads, so to speak, to know exactly what the problem is or what's causing the issue.
390
Two out of three people mistook the AI for a human therapist.
391
Does that mean that AI could, in some cases, actually replace a real person?
392
I'd say it's unlikely that more therapy slots will become available because of technology replacing certain aspects of the work.
393
Because there are always two arguments in this whole AI discourse.
394
It's about increasing efficiency and it's about supposedly improving quality.
395
But people rarely talk about putting more money into something just because AI can supposedly make it more efficient.
396
If AI chatbots can't really help when it comes to mental illness, at least not yet, then where can they help?
397
Perhaps as couples therapists, the way Daniel and Lucas are using one?
398
It's Daniel here.
399
Lucas and I are both here again, and I want to ask you to pose a few more questions that might help us make a decision,
400
to help us clarify what we want and to get things clear in our heads.
401
We've been given a whole list of questions to reflect on.
402
He made it so comprehensive, these questions are great.
403
He says we should both take 15 minutes to write down our answers.
404
Keywords are fine, then we take turns sharing them with each other.
405
Did writing it down help you?
406
It really helped.
407
What do you love about me even in this tough time?
408
The effort you put in to make our time together just unbelievably wonderful whenever we see each other.
409
I just love the energy you bring, making sure we have a great time.
410
What would I lose if we break up now?
411
The person I've spent the most time with over the last six months and who knows me so well.
412
The person I love and I love you.
413
Your answers to what I love am Anderen are strong, warm and full of awe.
414
This is a basis that many couples don't have.
415
The wounds are deep.
416
You notice that these wounds are not just stuck, but they break again, as soon as there is a distance.
417
As ChatGPT said, it's not unlikely that there will be more conflict.
418
It doesn't see how we interact on a personal level, how we touch.
419
How when I touch your hand I notice you have a little smile on your face.
420
A therapist would be.. probably be able to read both of us really well in that way.
421
Sometimes the AI just seems too optimistic, you know?
422
It doesn't dig deeper or ask critical questions or rubs salt in the wound.
423
We've realized how much we mean to each other and that we want to keep trying and hope for the best, right?
424
Very good.
425
Very good.
426
I've always been an optimist.
427
What stuck with me is ChatGPT saying
428
Chile would be an opportunity for us. Who knows?
429
Maybe Daniel will come visit me.
430
Right now, when it comes to mental health and counselling, AI chatbots can only play a supporting role.
431
The actual work still has to be done by people, that is, therapists and patients.
432
But when used correctly, AI can be a helpful tool for self -reflection.
433
Lots of studies show that just talking about things that are troubling you can have a positive effect.
434
But that doesn't mean I'd call it therapy.
435
Of course it's not pleasant when you come out of a grueling psychotherapy session where you've had to confront your shortcomings. Of course it's exhausting.
436
The subjective feeling of, oh, that was a nice conversation, or that conversation today was helpful,
437
that says absolutely nothing about the medium to long -term success of the therapy, nothing.
438
Even if AI can give answers comparable to those of human psychotherapists, I don't think it'll take our jobs away.
439
AI can respond by the book based on what it's been fed.
440
But our job is also to take a patient -specific and diagnosis -specific approach.
441
We don't impose a one -size -fits -all solution and say, here, I have a single right solution for you.
442
Patients should find the solutions that work best for them, and we should offer them help that enables them to help themselves.
443
You put me in an under a minute
444
Throw out every thought in my head
445
Till you only hear crickets
446
Put every face on me
447
That's under the sun
448
Until every feeling gets over You say, come on And I can talk
449
In any language
450
And I can know every answer
451
That's ever been known
452
But when everyone's together
453
Everyone's alone Cause we look so much better
454
When we're in the phone Thank you.
Download App
Everything you need to speak fluently
AI PronunciationScore every sentence
IPA PracticeMaster every sound
VocabularyBuild your word bank
Vocab GameLearn while playing
Same Channel
B21
BookTok and Gen Z - Are young people saving the book market? | DW Documentary
Daily ConversationWorking Adults
B2China's competition for living space | DW Documentary
Daily ConversationWorking Adults
B21
Cheap and toxic - The dark side of mass-produced fashion jewelry from China | DW Documentary
Daily ConversationWorking Adults
C13
Skin: Multitalented and underappreciated | DW Documentary
Daily ConversationWorking Adults
C12
Virtual love - How dangerous are AI relationships? | DW Documentary
Daily ConversationUniversity StudentsWorking Adults
B24
What it's like to be a working homeless person in Germany | DW Documentary
Daily ConversationWorking Adults
C111
Global renewables: Pioneering the energy transition | DW Documentary
IELTS PrepBusiness EnglishUniversity Students
B27
Father unknown? Life as a sperm donor's child | DW Documentary
Daily ConversationUniversity StudentsWorking Adults
B24
Our memory - A phenomenal storage system | DW Documentary
Daily ConversationUniversity StudentsWorking Adults
B22
The secret to concentration - How to increase your attention span | DW Documentary
Daily ConversationWorking Adults
C12
Junk food, sugar and additives - The dark side of the food industry | DW Documentary
Business EnglishDaily ConversationUniversity Students
C18
Can trees save the planet? | DW Documentary
IELTS PrepTOEFL PrepUniversity Students
C1Do we need to eat meat? | DW Documentary
Daily ConversationWorking Adults
C1Smart insects - The astonishing intelligence of bumblebees, wasps & co. | DW Documentary
IELTS PrepTOEFL PrepUniversity Students
B2Rivers under pressure - From drought and competing agendas | DW Documentary
IELTS PrepTOEFL PrepUniversity Students
B2Why India is one of the most dangerous places in the world for women | DW Documentary
University StudentsWorking Adults
FOR YOU
Recommend Video
C11
109 sentences
Teacher to Teacher: Is AI reshaping education for the better?
Daily ConversationUniversity StudentsWorking Adults
C14
71 sentences
How Psychotherapy Works
British AccentDaily ConversationUniversity StudentsWorking Adults
C12
182 sentences
We Let $200 GPT-Pro Think for 7 Days on Unsolved Math
Daily ConversationUniversity StudentsWorking Adults
C112
14 sentences
Would You Upload Your Mind?
Daily ConversationUniversity StudentsWorking Adults
C18
55 sentences
When AIs act emotional
Daily ConversationUniversity StudentsWorking Adults
C11
198 sentences
Jack Dorsey’s New Dessa AI Voice Sounds Scary Human
Daily ConversationUniversity StudentsWorking Adults
C12
171 sentences
Why the Arab World never united
Daily ConversationUniversity StudentsWorking Adults
B21
292 sentences
Learn German | German Podcast | Ep 211: Overthinking stoppen
Daily ConversationUniversity StudentsWorking Adults
B23
78 sentences
5 Things You Should Never Ask ChatGPT | Vantage with Palki Sharma
Business EnglishDaily ConversationUniversity Students
C13
85 sentences
AI Will End Philosophy, Physicist Claims
Daily ConversationUniversity StudentsWorking Adults
C11
171 sentences
What Does an AI Engineer Do Now? - Aparna Dhinakaran, Arize AI
Daily ConversationUniversity StudentsWorking Adults
Vocabulary and speaking notes for this lesson
This speaking lesson is built on the video “Artificial intelligence and psychotherapy”. This video has 454 sentences and 5241 words to shadow. The speech runs for 42:59. The speaker talks at a steady 122 words per minute, a comfortable pace for shadowing. 82% of the words are among the 3,000 most common in English; the rest is worth studying before you start.
Key vocabulary in this video
15 words from the video worth learning, with pronunciation and meaning:
| Word | Pronunciation | Meaning |
|---|---|---|
| therapist noun | /ˈθɛɹəpɪst/ | Someone who provides therapy, usually professionally. |
| anxiety noun | /æŋˈzaɪ.əti/ | An unpleasant state of mental uneasiness, nervousness, apprehension and obsession or concern about some uncertain event. |
| ward noun | /wɔɹd/ | The action of a watchman; monitoring, surveillance (usually in phrases keep ward etc.) |
| psychiatric adjective | /ˌsaɪ.kiˈæt.ɹɪk/ | Of, or relating to, psychiatry. |
| everyday adjective | /ˈɛvɹiˌdeɪ/ | Appropriate for ordinary use, rather than for special occasions. |
| assessment noun | /əˈsɛsmənt/ | The act of assessing or an amount (of tax, levy or duty etc) assessed. |
| helpful adjective | /ˈhɛlp.fl̩/ | Furnishing help; giving aid; useful. |
| illness noun | /ˈɪl.nəs/ | An instance (episode) of a disease or poor health. |
| psychological adjective | /ˌsaɪ.kəˈlɑː.d͡ʒɪ.kl̩/ | Of or pertaining to psychology. |
| experiment noun | /ɪkˈspɛɹ.ə.mənt/ | A test under controlled conditions made to either demonstrate a known truth, examine the validity of a hypothesis, or determine the efficacy or likelihood of… |
| diagnosis noun | /daɪəɡˈnəʊsɪs/ | The process of, or an instance of, identification of the nature and cause of a medical condition or illness. |
| symptom noun | /ˈsɪm(p)təm/ | A perceived change in some function, sensation or appearance of a person that indicates a disease or disorder, such as fever, headache or rash; strictly, a… |
| routine noun | /ɹuˈtin/ | A course of action to be followed regularly; a standard procedure. |
| comprehensive adjective | /ˌkɑm.pɹəˈhɛn.sɪv/ | Broadly or completely covering; including a large proportion of something. |
| psychology noun | /saɪˈkɑ.lə.d͡ʒi/ | The study of the human mind. |
Phrasal verbs you will hear
| Word | Pronunciation | Meaning |
|---|---|---|
| end up verb | To bring to a conclusion. | |
| break up verb | To break or separate into pieces. | |
| follow up verb | To take further actions remaining after an event; to continue, revisit, or persist; especially, to maintain communication or verify. | |
| get over verb | To overcome. | |
| go back to verb | To stem from; to originate in. | |
| look forward to verb | To anticipate, expect, or wait for, especially with a feeling of approval or pleasure; to be excited or eager to. | |
| run out verb | To use up or consume all | |
| work out verb | /ˌwɝk ˈaʊt/ | To calculate. |
Sentences worth repeating
Short, complete lines from the video that you can reuse in everyday conversation:
- What would I lose if we broke up?
- What are some things we can do?
- How much experience do you have with psychotherapy?
- How do you feel about that?
- I was like, is there something wrong with me?
Grammar in this video
The structures the speaker uses most, with the exact words from the video:
| Structure | In the video |
|---|---|
| Present perfect continuous have/has been + -ing — an action that started earlier and is still going on | We've been talking · I've been dealing |
| Present perfect have/has + past participle — a past action that still matters now | who've tried · have always tried · that's been |
| Passive voice be + past participle — the focus is on what happens, not who does it | aren't designed · could be delegated · are asked |
| Relative clauses who / which + clause — extra information about a person or thing | someone who listened · those who receive · say which was |
Pronunciation to watch
The speaker uses 151 contractions and reduced forms, such as I'm, I've, don't. Say them the short way, as you hear them.
- The “th” sounds: therapist /ˈθɛɹəpɪst/, breathe /bɹiːð/, psychotherapy /ˌsaɪkoʊˈθɛɹəpi/
- The “sh” and “zh” sounds: artificial /ˌɑː.tɪˈfɪʃ.əl/, emotion /ɪˈməʊ.ʃən/, passion /ˈpæʃən/, efficient /ɪˈfɪʃənt/, essentially /ɪˈsɛnʃəli/
- Long words — get the stress right: anxiety /æŋˈzaɪ.əti/, psychiatric /ˌsaɪ.kiˈæt.ɹɪk/, psychological /ˌsaɪ.kəˈlɑː.d͡ʒɪ.kl̩/, experiment /ɪkˈspɛɹ.ə.mənt/, comprehensive /ˌkɑm.pɹəˈhɛn.sɪv/
How to practise with this video
- Listen to the whole video once without speaking and note the words you do not know.
- Shadow it sentence by sentence at normal speed, repeating each one until your rhythm matches the speaker.
- Record yourself and compare with the original, paying attention to words like therapist, anxiety, ward.
What is the Shadowing Technique?
Shadowing is a science-backed language learning technique originally developed for professional interpreter training and popularized by polyglot Dr. Alexander Arguelles. The method is simple but powerful: you listen to native English audio and immediately repeat it out loud — like a shadow following the speaker with just a 1–2 second delay. Unlike passive listening or grammar drills, shadowing forces your brain and mouth muscles to simultaneously process and reproduce real speech patterns. Research shows it significantly improves pronunciation accuracy, intonation, rhythm, connected speech, listening comprehension, and speaking fluency — making it one of the most effective methods for IELTS Speaking preparation and real-world English communication.