Shadowing Practice: The Death of Educational Content on YouTube - Learn English Speaking with Video
Choose a practice mode
Each mode opens a full-screen practice room.
391 sentences
If sentences are too short or too long, click Edit to adjust them.
1
There's this brilliant KFC ad.
2
Welcome to Chicken Town.
3
Home of the imitators.
4
Because when you're on top, everybody wants a piece of you.
5
And lately we've been feeling a lot like KFC.
6
There's this thumbnail, this one, this...
7
What even is this?
8
Every thumbnail we invent, every title we construct, every animation we build, sometimes entire videos.
9
Countless AI slop channels relentlessly copy us.
10
And one day we decided, enough is enough.
11
So we investigated.
12
We looked into these channels and found who's likely behind them.
13
We joined a Discord community, paid 75 bucks for coaching
14
and subjected our fact checking team to an unhealthy amount of
15
AI misinformation spread to hundreds of thousands of people on topics like, well, the Holocaust.
16
Because all of this is endangering educational content on YouTube.
17
It's damaging for everyone in our niche.
18
It undermines trust in journalism.
19
And really, it's just so annoying.
20
They think that my voice is AI.
21
They even think that my face is AI generated.
22
The cost of production has gone down tremendously.
23
This is a new form of industrialized plagiarism and creative theft.
24
True crime, video essays, science, no space is safe.
25
We're in the golden era of soulless slop.
26
I think a couple of years ago, a 3D thumbnail was associated to good quality content, but now it's perceived as AI slop.
27
There's no wish to make something creative.
28
It's just a get-rich scheme.
29
AI is undoubtedly useful in many aspects of life and it's here to stay.
30
We use AI tools selectively in our own organization, for example to organize and skim through our research.
31
In rare cases, we use generated imagery as a basis for compositions in Photoshop if no comparable stock images are available.
32
But the rise of specifically generative AI has also made it very easy to mass produce content.
33
There's AI music on Spotify, terrible AI recipes and home decor on Pinterest.
34
Boomers are losing their mind on Facebook.
35
This is another account that I found that posts images like this.
36
It always seems to be a child sculpting something very impressive out of trash
37
and the people in the comments just eat this stuff up dude.
38
Great work, keep it up, great talent, nice one, thanks.
39
But one platform especially has become a dumping ground for AI slop.
40
More than 20% of the videos YouTube's algorithm shows to new users may now be slop.
41
The cost of production has gone down tremendously.
42
You can have content which is not so good or reasonable in quality, which is not competing against something
43
which is generated by a human who has spent a lot of time and investment and invested into acquiring his skills.
44
This is Dr. Akhil Bhattwush.
45
He's an associate professor at the University of Bath School of Management.
46
And then you have this thing which can generate precisely that at a value of cost.
47
And they're both competing.
48
There's something genuinely frustrating about clicking on a video and then realizing it's just another AI voice reading an AI script
49
feeding you AI visuals.
50
And we reached a full circle moment when we came across this video about the invasion of AI's slop made with AI.
51
The only comment reads, oh god, this video is AI too.
52
The creator gave it a heart.
53
And it's moments like these that make you believe the internet is really dead.
54
Around 100 people work on FERN, our German channel and our German podcast.
55
A dedicated team of journalists, art directors, producers, 2D and 3D animators and fact checkers.
56
All to bring you that weekly dose of something to watch during lunch.
57
But now, thanks to the power of AI, you can be Fern too.
58
You can wear our skin and talk in our voice and walk in our shoes.
59
So I'll take screenshots from Fern's video, paste them into chat GPT and simply say, regenerate this photo.
60
Most people think you need to be a professional 3D designer with years of experience to create videos like this.
61
Well, they're dead wrong.
62
I've cracked the code, he uses a Fern idea generator, a Fern script writer and a Fern image generator.
63
It's literally identical to what Fern uses and I would dare to say
64
that ours could be even better than what Fern is using here as an example.
65
And the best part, you don't have to be good at literally anything anymore.
66
You see, I don't have that budget available to me.
67
Neither do I have the skills to animate, do academic research, or I have that smooth voiceover.
68
But I do know how to use AI.
69
Some of it is just outrageous misinformation.
70
You've probably seen channels like Fern on YouTube where everything looks like it came out of Blender.
71
And when you first see it, you probably think that everyone spent weeks building it inside Blender.
72
And they actually didn't.
73
We actually did!
74
And it just doesn't end.
75
Yikes.
76
Imitation is flattery, we get that.
77
We've been inspired by the greats on this platform all our career.
78
But to us there's a difference between countless hours in After Effects to recreate Vox style maps
79
and asking a machine to just clone something.
80
We get why they choose to copy our work.
81
Edutainment is a good niche to target for slop gods.
82
Channels are often faceless.
83
The version of the animations are now replicable with AI, and people like to learn a little something while binging down a Hot Pocket.
84
The thumbnails don't rely on the creator's face, so you can fool people into thinking any considerable effort went into whatever your chain of tools spit out.
85
So of course, we're not the only edutainment channel having to deal with this.
86
So when I started making videos, I very much built my channel around 3D animations and high quality visuals.
87
And a few years ago, I think that was very unique and there wasn't a lot of that out there,
88
But now, that very style that I based my channel off has become kind of synonymous with AI slop.
89
And that's a challenge for me.
90
What I've noticed is that several people in my audience look at my 3D style of thumbnails
91
and they associate them to low effort content.
92
And my videos take a couple of months to research and produce, so I'm struggling a bit to convey that in the thumbnail, while keeping it attractive enough to get clicks.
93
I've seen videos take inspiration from my visual style before, which is fine by the way, but this was very different.
94
These visuals weren't just recreated in some animation software to look like mine, they were literally the same scenes with just a few things changed.
95
There's even math slop, which is just oh god.
96
Imagine trying to actually learn something on YouTube like it's 2017.
97
our thumbnails get ripped a lot and slightly changed by slop channels, likely using AI.
98
So we thought it would be funny to use AI to find matches.
99
Yannick, our data guy, because we actually have talented humans working here, set up a program to identify cases in which our thumbnails get lifted.
100
And there are so, so many.
101
We confronted several of the creators of these tutorials.
102
Some took down their videos afterward, others explained that they didn't earn money from their tutorials and others considered the tutorials free marketing for us.
103
One creator stated clearly he just wanted to make money and had only briefly used Fern as an example.
104
With regards to We Have Fern at Home, one AI channel stands out in particular.
105
Loom.
106
It's part of Loom Studio, a network of three channels.
107
Loom, Rocks and Ago.
108
In total they have more than 300,000 subscribers.
109
Ago and Loom seem to generate quite some revenue.
110
These channels regularly appropriate our thumbnails.
111
They rip them and then likely let AI change or add elements.
112
But sometimes they also seem to simply copy and recreate full videos of ours like this one.
113
Some of their content gets over a million views.
114
They also seem to take from others like MrBeast.
115
In many thumbnails MrBeast gets replaced with a 3D character.
116
In one instance a video with such a thumbnail got over 2 million views.
117
That could equate to 11 thousand dollars in revenue.
118
MrBeast even reacted to it on X.
119
Did they literally just swap me on my thumbnail for a red guy and get 1.8 million views?
120
What the hell lol?
121
But instead of being embarrassed, Loom Studio seems to be so proud of it, they even put it on their website.
122
MrBeast tweeted, we just kept going.
123
We confronted Loom about this and they told us they've never reused another creator's thumbnail.
124
Their content is supposedly created by their team from scratch.
125
But as if all that wasn't cringe enough, Loom Studio also offers a coaching program.
126
Their VIP course costs a casual $1499.
127
Included a blueprint plus weekly calls with our founders.
128
We review your channel, fix what's wrong and push you to your first 100k.
129
They seem to be pretty experienced.
130
They say they've built more than 100 faceless channels and generated over 100 million views.
131
And we can see why.
132
This company is run by a team of professionals.
133
A lion, a wolf and a tiger.
134
screams men building their AI company more than three predator mammals with sunglasses.
135
The wolf conjures AI pipelines from thin air, the one who makes the machine think, create and scale so humans don't have to.
136
Oh, how I hate having to be a human that thinks and creates.
137
Finally, the wolf can make it stop.
138
We dug deeper and found that Loom sits at the center of a small network of AI channels supporting each other.
139
Yannick created this map to show how these channels are linked together.
140
Whenever two channels collaborated on a video, a line connects them.
141
True Crime Vault pumps out fictional true crime stories, which is just a fantastic oxymoron.
142
It occasionally collaborates with other channels of the Loom network, so we suspect it could be a client of Loom Studio.
143
We asked the operator via email but didn't receive a reply.
144
The channels Ego, Loom, Rox and Loomicus all link to the same Instagram account of the lion Ardem.
145
Just like in nature, the lion seems to be the real king behind this company.
146
On Insta we find a profile with the same name, Ardem Eye, which has a link to Loom in his profile.
147
Ardem often posts pictures of his vacations, for example in Dubai.
148
There are luxury cars, fancy hotels, a big yard.
149
Chad, how do I say something is a fat f*ing cliché without saying it?
150
You could try something like, we're wandering straight into a well-worn narrative.
151
Okay, got it, got it, got it.
152
Yeah, yeah, yeah, sure.
153
Wow, yeah, we're wandering straight into a well-worn narrative here.
154
Our research team found out that Artem is probably from Ukraine.
155
We found an address in the north of the country.
156
He also seems to run a Ukrainian telegram channel about AI.
157
In the past, he seemingly used a different last name.
158
Of course, Luma are not the only ones doing business like that.
159
We want to dive deeper into the scene.
160
So we join a Discord channel.
161
YouTube faceless generally promotes a kind of hustle culture.
162
Not all of it is about AI, but quite a big part.
163
I've tried doing AI generated videos, but I think it's just easier to steal people's content as long as you react to the video.
164
AKA using AI voiceover, I think you can still get monetized.
165
Directly under that comment another user links our fentanyl video.
166
Where do you get the vids from?
167
Fern channel.
168
Can someone tell me how to get Fern?
169
Let me know cypher AI voice.
170
Hey, anyone here in the Fern Disrupt niche?
171
I feel bad for Fern and the Fern animators, lol.
172
Haha, why is that?
173
Because they spend hours animating 10 to 30 seconds of footage, and some random kid with AI can replicate their quality.
174
But can they really replicate our quality?
175
We decide to take a course to find out.
176
We aren't ready to spend one and a half thousand dollars to take the Alpha Animal coaching.
177
Luckily, AI with Gary's course comes at a bargain price of $75.
178
this video totally using AI-generated video and YouTube automation.
179
If you already know, there's a channel called Fern, and this YouTube channel is blowing up with cinematic 3D animations.
180
In their videos, they're using a lot of mannequin figures, and it's really effective for this kind of content.
181
So, if you're a beginner or still stuck on creating high-quality YouTube content, you need Gary Studio to speed up your workflow.
182
You also need someone with experience to teach and guide you in detail.
183
Just join my community.
184
So, using a fake identity, we join.
185
First, we receive an email that the payment was successful.
186
A few hours later, we get another email from Gary.
187
He writes that we can now join school with a monthly subscription.
188
We thought the $75 would unlock all materials.
189
But the $75 is for content access.
190
The school community requires a separate subscription, so both payments are required to join.
191
We reluctantly pay another $67 and finally get access.
192
We're in and uncovered a recipe for successful and captivating AI documentaries on YouTube.
193
Multiple shitty Excelists.
194
On top, AI with Gary provides a basic tutorial on how to produce AI videos.
195
From finding your niche and choosing a style, to word-for-word prompts for writing AI scripts or generating animations.
196
So far we just looked at the process.
197
But what about the results?
198
It's a pretty crucial aspect of edutainment after all.
199
How much does it eat you?
200
We bring in our head of fact checking Pascal
201
because we do have a team of fact checkers who go through every single one of our scripts and look for mistakes.
202
Pascal has years of experience in investigative journalism but now she faces the ultimate challenge.
203
Sit through hours of slop and fact check it.
204
We published this video on Fern in November 2025.
205
We sent a production crew to Scott Payne to let him tell his story.
206
Just boring ol' human activity.
207
Just weeks later, this video went online on a channel called Black Files.
208
Only seconds into the video and the opening scene is already misleading.
209
In the scene the video describes, the Klansmen did not wear these robes.
210
In this scene, the robes are incorrect.
211
They feature a red cross within a circle, not just a red cross.
212
And Pascal suspects that this entire scene is just made up by AI.
213
It directly contradicts the detailed sources she could find.
214
She also tried to fact check this video, but just conceded.
215
Quote, the language is so weird, it is hard to fact check.
216
Most sentences are not proper sentences.
217
It's hard to follow.
218
And yeah, yeah we get it.
219
This last example might be funny because it's bad, but factual errors and apparent hallucinations and content covering hate groups like the KKK are highly problematic.
220
We reached out to Blackfellz for comment
221
and they replied to Channel Defense's production process saying its videos
222
are created by a team using AI alongside in-house technology and I've reviewed for accuracy and originality.
223
They deny copying or infringing on third-party thumbnails
224
but acknowledge inaccuracies in the video we referenced and say they have since updated it.
225
Anyways, we can't help but wonder how these AI channels cover the Holocaust.
226
This video by Ego reached more than half a million people.
227
A true historical documentary about courage, moral failure and one escape that changed history forever.
228
Let's start with a visual highlight prisoners inexplicably collecting, shoveling and throwing a mountain of toys into a fireplace.
229
We asked a Holocaust expert about this and she told us the scene is obviously inaccurate.
230
The victims' belongings like toys were taken and collected but not piled up and burned in the crematoria.
231
So someone just made this up.
232
Maybe the creator, maybe the AI, who knows.
233
At one point the AI voice says, By 1944, his figure was staggering.
234
1,750,000 victims.
235
And the world did not know.
236
That was the secret of the system's success.
237
Secrecy.
238
Lies.
239
No one outside these walls knew the truth.
240
That figure of 1.75 million did circulate for a while and it did come from Rudolf Werber himself,
241
a man who escaped Auschwitz and whose story is covered in this true historical documentary.
242
But the video fails to mention that it does not reflect the current historical consensus.
243
Today historians estimate that around 1.3 million people were deported to Auschwitz between 1940
244
and 1945 and around 1.1 million were murdered there.
245
And it's so important to be accurate about this.
246
Holocaust deniers often exploit outdated or imprecise numbers to cast doubt on the Holocaust as a whole.
247
That's why videos covering this should clearly distinguish between historical estimates made at the time and the current academic consensus.
248
The phrase, the world did not know, is also very misleading.
249
In 1942, allied nations had already publicly declared that Germany was carrying out Hitler's so-called final solution to the Jewish question,
250
the mass murder of Jews.
251
Historians also believe many Germans had at least a rough idea of what was happening as information
252
and rumors about the crimes circulated.
253
And last but not least, we again reach a full circle AI bullshit moment when we stumble across this scene.
254
No one had ever escaped from Auschwitz to tell the world the truth.
255
That's incorrect, but also contradictory to Ego's own content.
256
The channel had already published a video about Vitol Pilecki, another man who infiltrated and escaped Auschwitz.
257
We covered that story ourselves.
258
And Pilecki, as well as multiple of his comrades, escaped or bought their freedom before Werber, delivering reports about Auschwitz to the outside world.
259
Maybe we shouldn't let the Wolf's machines do all the thinking and creating.
260
I think with a lot of these educational AI channels, the visuals are really only half the problem.
261
But a huge concern that I have is the quality of the information in those videos,
262
because you can be sure they're also using AI to generate all of their scripts.
263
There's one channel that recently got very popular, Chloe versus History, about a lady vlogging in the past and people love it.
264
But if you actually look at what she's saying, if you look at the details, if you look at the whole story, almost everything is wrong.
265
And especially with educational content, I think there's this idea that people trust the content and they walk away being informed.
266
but actually they're walking away being less informed about certain topics and I think that's really bad.
267
These AI channels don't seem to care about the quality or accuracy of their content.
268
Their goal seems to be simply to publish as many videos as possible to make as much money as possible.
269
Oh and I almost forgot we actually spoke with a person familiar with the production process behind Loom's channels.
270
They told us that they indeed use AI and that two people had to produce one script per day each.
271
We confronted Loom and sent them many questions.
272
In their response they wrote half of this is totally fictional and 30% wrong.
273
They also told us that their audience doesn't watch Fern, which we are keen to believe.
274
They claim that 30 people work at their company with AI being used as a tool
275
and lots of manual labor still involved.
276
They were in disbelief of our apparent hate of them.
277
Lastly, life is like a boomerang.
278
Okay.
279
There's no wish to make something creative.
280
It's just a get-rich-scheme." And it's not just teenagers and Discord channels
281
and cool guys who keep going even when MrBeast tweets driving the slop epidemic.
282
Various AI companies seem to know exactly who their target audience is and how their products are used.
283
Higgsfield offers a variety of generative AI tools and recently they uploaded this tutorial on their channel.
284
I used Higgsfield AI and Claw to build a $39,500 a month faceless channel.
285
I just found a faceless YouTube channel that's making almost $40,000 every single month.
286
And it made me wonder, can I completely recreate this channel and build the exact same style of videos using AI?
287
The 40k a month educational channel he found is Brightside.
288
In case you're not familiar with Brightside, the incredible irony might be lost on you.
289
Brightside is one of the OG slop factories before there was AI as we know it.
290
Terrible clickbait, unhinged content pumped out every few hours.
291
It's just colorful thumbnails, interesting clickable titles.
292
It's.. they make things.. they make things, that's about it.
293
Have you never heard of Brightside?
294
Good for you.
295
Brightside just straight up lies or just uses these like fear-mongering sensationalist tactics.
296
Yet this is the type of content Higgsfield presents as a great opportunity to make bank with their tools.
297
Educational content appeals to everyone and has huge audience retention, which is exactly what drives you monetization.
298
In this brave new world, educational videos obviously aren't made to educate, they're simply the most lucrative niche to make as much money as possible.
299
And we can see how well that logic plays out just a few minutes later
300
when he uses AI to create a video about Pompeii.
301
We could spend minutes explaining how much is wrong with these glass jars, but simply put, they just didn't exist in this form in Pompeii in 79 AD.
302
Now, you might wonder, who would possibly want to slurp up this type of slop?
303
Well, the people at Hicksfield.
304
I think this video is very interesting.
305
This is the kind of stuff I watch right before going to bed.
306
Really puts the old saying that you should make stuff you would want to watch yourself into perspective.
307
Next, the tutorial demonstrates how to make even more videos, fully generated with basically no human input.
308
Okay, one video is cool, but we need more.
309
So I'm just gonna type in, make me, two more videos.
310
This is great news for anyone who wants to get paid by YouTube without actually making the content themselves.
311
Let's launch our content machine that will bring in money.
312
During a gold rush, sell shovels.
313
Eleven Labs offers AI voice tools.
314
And they have their very own tutorial on how to make AI B-roll that's just like Fern.
315
And what do you do once you have your slop visuals in place?
316
Add a slop voice.
317
As he began his speech, there was a shift in the room.
318
You could feel it, like the start of something important.
319
During a gold rush, sell shovels.
320
We sent both Hicksfield and Eleven Labs a request for comment, but didn't receive a reply.
321
Let's zoom out.
322
In the best case scenario, these AI productions just result in a few ridiculous visual mistakes.
323
Worst case, they lead to misinformation about some of humanity's darkest times.
324
And how would viewers possibly know that what they're watching is wrong?
325
Pascal, seasoned investigative journalist and our head of fact checking wrote this after reviewing some of the videos.
326
It is incredibly hard to verify things that never actually existed.
327
Real people usually leave some kind of trace or reference point, but AI can just completely make things up.
328
If it's already difficult for me, I don't think average users can realistically deal with this.
329
After just checking half a video, my head was genuinely spinning.
330
Luckily many people instinctively distrust AI content, but that leads to a loss of trust in real creators too.
331
I get comments on my videos all the time where people think that the 3d animations are made using AI, they think that my voice is AI,
332
they even think that my face is AI generated.
333
People get so anxious about slop they start to suspect it everywhere.
334
If YouTube lets this kind of content go unchecked and allows its creators to profit from it, in the long run it could damage the platform's reputation and value proposition.
335
So what is YouTube doing about it?
336
Already in 2025 the platform updated its channel monetization policies.
337
They renamed repetitious content to inauthentic content.
338
Low effort, mass produced uploads shouldn't make money anymore.
339
And in 2026 they updated them again with their unsatisfying or off-putting content policy.
340
YouTube now targets content that feels interchangeable, emotionally manipulative or mimics existing formats.
341
Great news for us, hopefully.
342
YouTube also pushed towards disclosure and labeling.
343
SynthID is Google DeepMind's watermarking system.
344
It embeds invisible, irremovable markers into AI-generated content.
345
Major generative AI companies such as OpenAI, NVIDIA, Kakao and Eleven Labs have adopted the watermark.
346
With 11 labs on board, more AI-generated voices can now be automatically detected by YouTube.
347
In May, the platform announced they would start automatically tagging AI-generated or AI-edited videos.
348
The labels will be clearly visible under videos or in shorts.
349
The goal is to make AI content easier to spot, even if creators don't disclose it.
350
But of course, slop factories will try to get around the tag in demonetization.
351
Hicksfield has them covered.
352
And for those of you who are worried about AI content getting demonetized, YouTube has no problem with AI.
353
It just blocks people who use it to make lazy low-quality spam.
354
This is an unironic statement in a video demonstrating how to automatically pump out bright side style content within 20 minutes.
355
We also asked YouTube about Slop directly and shared our concerns.
356
And they actually got back to us saying that tackling Slop is a key priority for them.
357
They also said this
358
YouTube's approach to information quality has always focused on promoting high-quality content rather than fact-checking individual videos at scale,
359
but under our misinformation policies we remove misleading or deceptive content with serious risk of egregious harm.
360
We do believe YouTube understands what's at stake and that the platform needs to crank up the heat even further.
361
Many years ago at the start of our career, YouTube was drowning in clickbait videos full of stock footage.
362
That changed over time as the platform managed to measure and reward retention and watch time.
363
Deceiving viewers didn't work that well anymore because YouTube would punish videos when people would click off fast.
364
Optimizing for retention leads to its own problems, but generative AI feels far more threatening to overall content quality, at least for edutainment.
365
This is the next thing harming the longevity of our ecosystem.
366
Personally, I increasingly find myself scrolling through my recommendations, irritated at every piece of slop served to me.
367
It makes me want to open the app less.
368
And that can't be good.
369
Contrary to some channels on YouTube, our 3D animations are made by humans.
370
This is one of the artists who regularly puts in countless hours in creating them for you.
371
Hi, I'm Leonard.
372
When I'm creating scenes, I can easily spend a whole day sitting at my computer.
373
So shoutout to Flexisport for sponsoring this video and sending me this.
374
One button and I can switch from sitting to standing.
375
Creating a scene like this roughly comes down to three steps.
376
Step 1, modeling.
377
It's similar to the real thing.
378
I stack simple cubes to block out the desk, then rest the set with a plant, computer and a display, the usual clutter.
379
Step 2, animating.
380
This desk actually moves, so I have to animate the legs sliding into each other, with a little ease out, so it doesn't feel robotic.
381
It's a three-stage frame, so it adjusts to you, no matter if you're 5'2 or 6'5.
382
Step 3.
383
Shading.
384
It looks a bit bland still.
385
I've got the white oak, but if you want something even more premium, they have great options for 100% solid wood desktops.
386
And it's not just a work desk.
387
It can be a gaming station, a dining table, or where all your stuff ends up living.
388
Mine's basically all three and even when I completely load it up, it's still really solid and stable.
389
If you want to upgrade your setup use code FERN50 for $50 off the E7 Pro.
390
And if you're not sure, Flexispot offers 30-day risk-free returns and a 15-year warranty.
391
Link is in the description.
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
B22
The Only Nazi Hitler Feared
Daily ConversationWorking Adults
B21
The Hunt for the Charlie Kirk Shooter
Daily ConversationWorking Adults
B21
Hitler's Insane Train
Daily ConversationWorking Adults
B21
How Were the Pyramids Actually Built?
Daily ConversationWorking Adults
C12
The Mutated Animals of Chernobyl
Daily ConversationWorking Adults
C11
The Dirty Business Behind Caffeine
Daily ConversationUniversity StudentsWorking Adults
B2How Iran’s Leader Was Killed
Daily ConversationWorking Adults
B2The Hunt for America's Smartest Killer
Daily ConversationWorking Adults
B2The Unsolved Murder in Room 1046
University StudentsWorking Adults
B2The World's Deadliest Child
IELTS PrepTOEFL PrepHigh School Students
FOR YOU
Recommend Video
B296 sentences
This Scriptless Speaking Style Is Quietly Blowing Up on YouTube
Business EnglishUniversity StudentsWorking Adults
B21
154 sentences
The 7 Trillion AI Gamble Is Failing. Big Tech is TRAPPED Right NOW.
Business EnglishDaily ConversationUniversity Students
B2162 sentences
Modrinth Just BANNED AI Mods
Business EnglishUniversity StudentsWorking Adults
B22
95 sentences
FYI - Chat series with EU Ambassador: Episode 4 - Filter Your Internet
Business EnglishDaily ConversationUniversity Students
B214
211 sentences
AI Ads Are Out of Control
Business EnglishDaily ConversationUniversity Students
B23
245 sentences
How To Become Dangerously Self-Educated (with AI)
Business EnglishUniversity StudentsWorking Adults
B22
152 sentences
Conan Learns About His YouTube Audience | Conan O'Brien Needs A Friend
Business EnglishDaily ConversationUniversity Students
B22
120 sentences
How To Actually Monetize Claude (4 Ways That Work In 2026)
Business EnglishUniversity StudentsWorking Adults
B22
278 sentences
Maybe we were wrong
Business EnglishUniversity StudentsWorking Adults
B22
79 sentences
AI is flooding the internet—here’s what I’m protecting
American AccentBusiness EnglishUniversity StudentsWorking Adults
B24
180 sentences
Is AI Doom Real... or Just Good Marketing?
Business EnglishUniversity StudentsWorking Adults
B24
168 sentences
Rage Bait RUINED The Internet.
Business EnglishDaily ConversationUniversity Students
Vocabulary and speaking notes for this lesson
This video has 391 sentences and 4741 words to shadow. The speech runs for 28:10. The speaker talks at a natural 168 words per minute, close to everyday conversation. Only 78% of the words are among the 3,000 most common in English, so the vocabulary is demanding.
Key vocabulary in this video
15 words from the video worth learning, with pronunciation and meaning:
| Word | Pronunciation | Meaning |
|---|---|---|
| generate verb | /ˈd͡ʒɛn.ə.ɹeɪt/ | To bring into being; give rise to. |
| creator noun | /kɹiːˈeɪ̯təː/ | Something or someone which creates or makes something. |
| script noun | /skɹɪpt/ | A writing; a written document. |
| niche noun | /nɪt͡ʃ/ | A cavity, hollow, or recess, generally within the thickness of a wall, for a statue, bust, or other erect ornament. |
| lab noun | /læːb/ | A Labrador retriever. |
| tutorial noun | /ˌtjuːˈtɔːɹɪəl/ | A self-paced learning exercise; a lesson prepared so that a student can learn at their own speed or convenience. |
| desk noun | /dɛsk/ | A table, frame, or case, in past centuries usually with a sloping top but now usually with a flat top, for the use of writers and readers. It often has a… |
| wolf noun | /wʊlf/ | Canis lupus; the largest wild member of the canine subfamily. |
| lion noun | /ˈlaɪən/ | A big cat, Panthera leo, native to Africa, India and formerly much of Europe. |
| countless adjective | /ˈkaʊntləs/ | Synonym of uncountable. |
| retention noun | /ɹɪˈtɛnʃən/ | The act of retaining or something retained. |
| confront verb | /kənˈfɹʌnt/ | To stand or meet facing, especially in competition, hostility or defiance; to come face to face with. |
| recreate verb | /ˈɹɛk.ɹi.eɪt/ | To give new life, energy or encouragement (to); to refresh, enliven. |
| topic noun | /ˈtɑpɪk/ | A subject; a theme; a category or general area of interest. |
| rush verb | /ɹʌʃ/ | To hurry; to perform a task with great haste, often not properly or without thinking carefully. |
Phrasal verbs you will hear
| Word | Pronunciation | Meaning |
|---|---|---|
| find out verb | To discover, as by asking or investigating. | |
| get over verb | To overcome. | |
| go down verb | To descend; to move from a higher place to a lower one. | |
| walk away verb | To withdraw from a problematic situation. | |
| carry out verb | /ˈkæ.ɹi aʊt/ | To hold while moving it out. |
| come across verb | To change sides; to cross over to work for the opposition. | |
| come down to verb | To reach by moving down or reducing. | |
| end up verb | To bring to a conclusion. |
Pronunciation to watch
The speaker uses 59 contractions and reduced forms, such as don't, they're, didn't. Say them the short way, as you hear them.
- The “th” sounds: thin /ˈθɪn/, smooth /smuːð/, mathematics /mæθ(.ə)ˈmæt.ɪks/, theft /θɛft/, threaten /ˈθɹɛt.n̩/
- The “sh” and “zh” sounds: retention /ɹɪˈtɛnʃən/, rush /ɹʌʃ/, associate /əˈsoʊʃi.ət/, animation /ˌæn.ɪˈmeɪ.ʃən/, subscription /səbˈskɹɪpʃən/
- Long words — get the stress right: associate /əˈsoʊʃi.ət/, documentary /ˌdɒk.jʊˈmɛn.tɹi/, accuracy /ˈæk.jɚ.ə.si/, animation /ˌæn.ɪˈmeɪ.ʃən/, genuinely /ˈd͡ʒɛn.ju.ɪn.li/
How to practise with this video
- Listen to the whole video once without speaking and note the words you do not know.
- Start at 0.75× speed, shadow it sentence by sentence, then go back to normal speed once it feels easy.
- Record yourself and compare with the original, paying attention to words like generate, creator, script.
Grammar in this video
The structures the speaker uses most, with the exact words from the video:
| Structure | In the video |
|---|---|
| Present perfect have/has + past participle — a past action that still matters now | has gone · has also made · has become |
| Passive voice be + past participle — the focus is on what happens, not who does it | was associated · is generated · We've been inspired |
| Relative clauses who / which + clause — extra information about a person or thing | content which is · thing which can · before, which is |
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.