シャドーイング練習: Will AI Take Your Job in the Next 10 Years? Wrong Question | Vinciane Beauchene | TED - YouTubeで英語スピーキングを学ぶ
上級
シャドーイング コントロール
0% 完了 (0/135 文)
Back in the 50s, Alan Turing came up with an idea.
⏸ 一時停止中
再生速度:
リピート回数:
待機モード:
字幕同期:0ms
すべての文
135 文
1
Back in the 50s, Alan Turing came up with an idea.
0:04.26 – 0:08.38 (4.1s)
2
If you couldn't tell if you were talking to a machine or a human, it meant the machine must be intelligent.
0:09.26 – 0:17.54 (8.3s)
3
He coined the Turing test.
0:17.58 – 0:19.34 (1.8s)
4
Today, most chatbots pass the test easily.
0:20.38 – 0:24.14 (3.8s)
5
But here's the catch.
0:25.14 – 0:26.98 (1.8s)
6
I believe the test was wrong.
0:27.02 – 0:29.10 (2.1s)
7
Because talking isn't what's going to change the world.
0:30.26 – 0:34.50 (4.2s)
8
Doing is. That's why I ask a slightly different question to the leaders I work with.
0:34.54 – 0:41.90 (7.4s)
9
On a daily basis, my role is to reshape organizations, trying to find the right mix of strategy, tech and talent.
0:41.90 – 0:51.66 (9.8s)
10
And my obsession is to make sure that talents do not get out of the equation.
0:52.30 – 0:57.78 (5.5s)
11
So the question I ask my clients is: if an AI could take over all of your team's tasks, who would you keep and why?
0:58.86 – 1:11.30 (12.4s)
12
That question is strategic.
1:11.34 – 1:13.78 (2.4s)
13
And the answer matters to me, not just intellectually, but because I have two daughters at home.
1:13.78 – 1:20.54 (6.8s)
14
They are five and nine.
1:20.58 – 1:22.34 (1.8s)
15
And right now, as you can see, they feel invincible.
1:22.38 – 1:26.66 (4.3s)
16
But I keep wondering: What is the world of work they will step into?
1:26.66 – 1:32.42 (5.8s)
17
We need to build a future where humans matter more, not less.
1:32.46 – 1:37.38 (4.9s)
18
Now let me try to illustrate how this is playing out in the field.
1:38.30 – 1:42.50 (4.2s)
19
A consumer goods client of mine is all in on AI.
1:43.50 – 1:47.74 (4.2s)
20
They didn't want to just deploy the next algo.
1:48.62 – 1:51.86 (3.2s)
21
They wanted to rethink the selling process itself.
1:51.86 – 1:56.22 (4.4s)
22
The trigger was agents.
1:56.26 – 1:58.66 (2.4s)
23
Have you heard about agents?
1:59.78 – 2:01.86 (2.1s)
24
They are the latest generation of AI: more autonomous, able to connect across systems, to plan, to take action, to learn, to adapt.
2:01.86 – 2:12.78 (10.9s)
25
The James Bond of AI.
2:12.78 – 2:14.58 (1.8s)
26
And applied to the selling process, you get an agent that is able to target the customer, make recommendations, negotiate, close the deal.
2:15.70 – 2:26.90 (11.2s)
27
All this with no human intervention.
2:26.90 – 2:29.50 (2.6s)
28
A fully autonomous sales engine.
2:30.38 – 2:33.42 (3.0s)
29
And it was technically feasible.
2:34.66 – 2:36.78 (2.1s)
30
But then an exec asked, "Hmm, if the machine does all of this, then what remains for humans?" This cracked everything open because when we looked deeper at their most loyal customers, we saw they weren’t sticking around because of prices or products but because of how the sales rep made them feel.
2:37.78 – 3:02.74 (25.0s)
31
So we flipped the model around.
3:03.74 – 3:06.34 (2.6s)
32
Humans were no longer going to be about pushing products.
3:06.38 – 3:10.26 (3.9s)
33
They were going to be about building relationship, belonging, loyalty.
3:10.30 – 3:15.90 (5.6s)
34
Very concretely, this meant new skills, new incentives, a very different mindset.
3:17.22 – 3:24.26 (7.0s)
35
Well it changed everything.
3:24.30 – 3:26.22 (1.9s)
36
But it worked.
3:26.26 – 3:28.90 (2.6s)
37
Because in the age of AI, human value isn't gone.
3:28.90 – 3:33.18 (4.3s)
38
It's just moved.
3:33.22 – 3:34.46 (1.2s)
39
Now I’m not talking about copilots anymore.
3:35.74 – 3:38.46 (2.7s)
40
For a while the narrative has been AI will augment us, not replace us.
3:38.50 – 3:45.06 (6.6s)
41
Well this is not where the tech is going today.
3:45.10 – 3:47.94 (2.8s)
42
And I believe we have real hard work to do if we want this narrative to stay true.
3:48.82 – 3:53.46 (4.6s)
43
So I'll say a few words about what I think needs to be done in a second.
3:55.50 – 3:59.74 (4.2s)
44
But first, let me tackle three myths that I think are holding us back.
3:59.74 – 4:04.58 (4.8s)
45
I call them "head in the sand" ideology.
4:05.82 – 4:09.02 (3.2s)
46
Number one.
4:10.26 – 4:12.02 (1.8s)
47
"All of this is overblown. We'll adapt." Yes. We've adapted to electricity, the industrial revolution, the internet.
4:12.06 – 4:23.82 (11.8s)
48
But we've done so on the back of generations that did not have the training nor the time to adapt.
4:23.82 – 4:29.58 (5.8s)
49
And in the case of this revolution, time is of the essence.
4:30.54 – 4:34.70 (4.2s)
50
You may think you have time because agents are just emerging.
4:35.66 – 4:39.10 (3.4s)
51
And it's a fact.
4:39.14 – 4:40.46 (1.3s)
52
Our research shows that today only 13 percent of companies have embedded agents in their workflows.
4:40.50 – 4:47.78 (7.3s)
53
But tech moves exponentially.
4:49.46 – 4:51.82 (2.4s)
54
Humans, they crawl linearly.
4:51.82 – 4:54.62 (2.8s)
55
If you don't prepare now, you'll struggle to keep up.
4:54.62 – 4:58.46 (3.8s)
56
And I'm not talking about science fiction.
4:59.26 – 5:01.94 (2.7s)
57
I'm not talking about AGI, artificial general intelligence, this moment where AI will be smarter than us.
5:01.94 – 5:09.98 (8.0s)
58
I'm referring here to ACI, artificial capable intelligence.
5:10.02 – 5:16.66 (6.6s)
59
The moment when AI will be able to take on ambiguous, complex goals with minimal oversight.
5:17.58 – 5:26.90 (9.3s)
60
And while AGI is speculative, ACI is a deadline.
5:26.90 – 5:32.62 (5.7s)
61
While we spend hours debating about superintelligence and consciousness, we miss the milestones that ACI is meeting with increasing frequency.
5:32.62 – 5:43.58 (11.0s)
62
ACI will change how work is done and by whom.
5:44.62 – 5:48.10 (3.5s)
63
Let's shape it, not wait and see.
5:49.06 – 5:51.50 (2.4s)
64
Now myth number two.
5:52.66 – 5:54.02 (1.4s)
65
"Soft skills are our sweet spot." Yes, it's lovely to believe that empathy, creativity are uniquely ours.
5:54.86 – 6:05.46 (10.6s)
66
But evidence says otherwise.
6:06.66 – 6:09.86 (3.2s)
67
More and more humans like to interact with AI because they feel it's more empathic.
6:09.86 – 6:16.90 (7.0s)
68
And why not?
6:18.46 – 6:19.74 (1.3s)
69
I mean, AI doesn't get tired, doesn't get cranky, doesn't judge you.
6:19.74 – 6:28.06 (8.3s)
70
So this moat we thought was ours, it's shrinking.
6:28.74 – 6:32.62 (3.9s)
71
And we need to stop asking what AI can't do and focus on where humans make a difference and why.
6:33.22 – 6:41.14 (7.9s)
72
At this stage, I'm sure you would love me to come up with the list of human qualities that will remain ours forever.
6:43.78 – 6:52.34 (8.6s)
73
But my point is, there is no universal list.
6:53.30 – 6:55.82 (2.5s)
74
Each company needs to figure it out based on its strategic positioning.
6:56.54 – 7:02.34 (5.8s)
75
This is hard, uncomfortable work, but it's work that you as leaders need to take on.
7:02.38 – 7:09.06 (6.7s)
76
Now myth number three.
7:10.34 – 7:11.74 (1.4s)
77
My preferred one. I'm French.
7:13.06 – 7:14.94 (1.9s)
78
"We need to protect jobs." Yes. I see where this one is coming from.
7:14.94 – 7:21.78 (6.8s)
79
Today, 41 percent of employees believe that their job will vanish in the next decade because of AI.
7:21.78 – 7:29.42 (7.6s)
80
But protecting jobs is like anchoring a boat in a storm.
7:30.62 – 7:34.78 (4.2s)
81
Jobs are fixed.
7:34.78 – 7:36.14 (1.4s)
82
The human potential to grow and adapt, on the other hand, is not.
7:37.06 – 7:40.94 (3.9s)
83
This is where we need to invest.
7:41.62 – 7:44.06 (2.4s)
84
The challenge is our organizations are not geared for that today.
7:44.10 – 7:48.02 (3.9s)
85
Org charts are static.
7:48.06 – 7:49.86 (1.8s)
86
Career paths are narrow.
7:49.86 – 7:51.78 (1.9s)
87
Training is occasional.
7:51.78 – 7:53.62 (1.8s)
88
This system will fall apart the day that the boundaries of jobs start melting away fast.
7:53.62 – 7:59.90 (6.3s)
89
So what do we need to do?
8:01.18 – 8:03.94 (2.8s)
90
Let me take you to an ideal company.
8:05.62 – 8:08.38 (2.8s)
91
Not a theoretical one, just the blend of the boldest clients I've worked with.
8:08.42 – 8:13.66 (5.2s)
92
First, they don't start with tech.
8:14.82 – 8:17.42 (2.6s)
93
They start with strategy.
8:18.62 – 8:20.02 (1.4s)
94
They focus on the outcomes that truly differentiate them on the market.
8:21.14 – 8:25.70 (4.6s)
95
They understand how agents will allow them to deliver against those outcomes in totally different ways.
8:26.38 – 8:33.50 (7.1s)
96
And they look at where people still make a difference for the better.
8:34.42 – 8:38.66 (4.2s)
97
As you can see, this is not incremental redesign of your operating model.
8:39.66 – 8:44.98 (5.3s)
98
It's radical AI-first reinvention.
8:45.02 – 8:48.34 (3.3s)
99
And we did this work for an industrial goods client of mine.
8:49.30 – 8:53.18 (3.9s)
100
Imagine having to go through 50 “hack a future” workshops, looking at how AI is going to disrupt each of your businesses, each of your function.
8:53.22 – 9:03.50 (10.3s)
101
Comfortable? It is not.
9:03.54 – 9:05.34 (1.8s)
102
But it allowed the leaders to align on a vision of where agents win, people matter and how best to pair them.
9:06.26 – 9:14.34 (8.1s)
103
Now once you have this vision, you want to translate it into a workforce model.
9:15.78 – 9:21.30 (5.5s)
104
How many people do I need?
9:21.34 – 9:23.26 (1.9s)
105
With what skills?
9:23.30 – 9:25.34 (2.0s)
106
No more guesswork, just informed, intentional reinvention.
9:25.38 – 9:30.34 (5.0s)
107
A multiyear skills forecast.
9:30.38 – 9:33.10 (2.7s)
108
And this is something we built for a consumer goods client that was facing a massive challenge.
9:33.14 – 9:39.42 (6.3s)
109
Imagine having to reformulate your entire product portfolio while keeping the leadership and innovation.
9:39.46 – 9:47.54 (8.1s)
110
Of course, AI unlocked the productivity that was required, but the work was much deeper.
9:47.58 – 9:54.10 (6.5s)
111
They needed to reinvent the role of the researcher.
9:54.14 – 9:57.14 (3.0s)
112
From chemist to data-driven biologist, from solo expert to multifunctional teammates.
9:58.06 – 10:05.74 (7.7s)
113
And they made it happen because they mapped very precisely the future skills that they needed, and they built a very effective upskilling and mobility engine.
10:05.74 – 10:15.50 (9.8s)
114
Then you want to publicly commit to taking your talents to their fullest potential.
10:17.10 – 10:24.02 (6.9s)
115
Now I know what you're going to tell me.
10:25.06 – 10:27.02 (2.0s)
116
Vinciane, why would we invest in talent if an AI can do their job faster, cheaper and without complaining?
10:27.06 – 10:37.38 (10.3s)
117
Well because the day that interacting with an AI becomes the new norm, a commodity, the interaction with humans is going to take an entire new meaning.
10:38.78 – 10:50.02 (11.2s)
118
Trust, authenticity, accountability.
10:50.62 – 10:54.42 (3.8s)
119
Those are the values we will anchor on.
10:55.46 – 10:57.86 (2.4s)
120
So the smartest companies will invest in talent.
10:58.94 – 11:03.22 (4.3s)
121
Not only tech talent, all talent.
11:03.26 – 11:06.14 (2.9s)
122
Not once, but systematically.
11:06.18 – 11:09.34 (3.2s)
123
And they will protect time to learn.
11:09.38 – 11:13.14 (3.8s)
124
Because today, while freelancers spend on average four hours per week learning, employees spend none.
11:13.18 – 11:22.74 (9.6s)
125
So no, the future isn't about being more human.
11:22.74 – 11:26.26 (3.5s)
126
It's about building the systems that will allow humans to do what matters most.
11:26.30 – 11:31.10 (4.8s)
127
This is not a story about job loss.
11:31.86 – 11:34.26 (2.4s)
128
It is a story about human differentiation.
11:35.22 – 11:37.86 (2.6s)
129
AI will keep on climbing. That is not up to us.
11:39.34 – 11:42.74 (3.4s)
130
But how fast we climb with it, that is up to us.
11:42.74 – 11:46.42 (3.7s)
131
So we need to stop asking: Will there still be jobs for humans?
11:47.30 – 11:52.30 (5.0s)
132
And focus on answering: What do we want humans to be best at?
11:52.34 – 11:57.66 (5.3s)
133
Because in the age of AI, being human isn't a fallback, it's a practice.
11:59.14 – 12:06.54 (7.4s)
134
Let's make it exceptional.
12:06.58 – 12:08.82 (2.2s)
135
Thank you. (Applause)
12:10.14 – 12:16.86 (6.7s)
コンテキストと背景
このTEDトークでは、Vinciane BeaucheneがAIの影響と人間の仕事の未来についての見解を示しています。彼女は、過去の技術革新と同様に、現在の人工知能の発展が私たちの働き方を一変させる可能性があると述べています。特に彼女は、AIが普及する中で人間の役割を再定義する必要性について強調し、人間の価値がどのように変化しているかに注目しています。
日常コミュニケーションのためのトップ5フレーズ
- 「AIはすべての仕事を奪うのか？」 - AIに対する恐れを表現する一般的な疑問。
- 「人間の価値は移動するだけだ。」 - AI時代における人間の役割の変化についての核心をついた言葉。
- 「顧客との関係構築が最も重要である。」 - セールスにおける人間の接触が持つ価値。
- 「新しいスキルとマインドセットが必要だ。」 - 変化に適応するための必要性を強調。
- 「時間が重要だ。」 - 技術の進化の速さに対する警告。
ステップバイステップのシャドーイングガイド
このビデオは、文章の速さや語彙が高度であるため、英語の発音を良くするためには少し時間がかかるかもしれません。以下の手順を参考に、シャドースピーチの練習を行ってみてください。
- ビデオを視聴する: 最初に全体を通して視聴し、内容を理解します。この段階で何度も繰り返し聴くことが効果的です。
- 重要なフレーズをメモする: 上記のトップ5フレーズを含め、心に残るフレーズをリストアップします。
- フレーズを分解する: 各フレーズを短いセクションに分け、それぞれに集中します。文ごとに停止し、自分の発音を確認しましょう。
- シャドーイングを行う: ビデオを再生し、スピーカーの後を追って話します。最初はゆっくりと、徐々にスピードを上げてみましょう。
- 録音して確認する: 自分の声を録音し、実際の発音と比較します。改善が必要な部分を特定し、練習を重ねてください。
このプロセスを通じて、YouTubeで英語学習をする際に重要なスキルが身につき、IELTSスピーキング対策にも役立つでしょう。定期的に練習をすることで、より自信を持って英語を話せるようになります。
シャドーイングとは？英語上達に効果的な理由
シャドーイング（Shadowing）は、もともとプロの通訳者養成プログラムで開発された言語学習法で、多言語習得者として知られるDr. Alexander Arguelles によって広く普及されました。方法はシンプルですが非常に効果的：ネイティブスピーカーの英語を聞きながら、1〜2秒の遅延で声に出してすぐに繰り返す——まるで「影（shadow）」のように話者を追いかけます。文法ドリルや受動的なリスニングと異なり、シャドーイングは脳と口の筋肉が同時にリアルタイムで英語を処理・再現することを強制します。研究により、発音精度、抑揚、リズム、連音、リスニング力、そして会話の流暢さが大幅に向上することが確認されています。IELTSスピーキング対策や自然な英語コミュニケーションを目指す方に特におすすめです。
ShadowingEnglishでの効果的な学習方法
- 動画を選ぶ: 自然で明瞭な英語が使われているYouTube動画を選びましょう。TED Talks、BBC News、映画のシーン、ポッドキャスト、IELTS模範解答などが最適です。URLをコピーして検索バーに貼り付けてください。短い動画（5分以内）や、自分が本当に興味を持てるテーマから始めるのがコツです。
- まず聞いて内容を理解する: 最初は1倍速でただ聞くだけにしましょう。まだ繰り返す必要はありません。文の意味を理解し、話者がどのように単語を強調し、音を繋げ、間を取っているかに注目してください。内容を把握してからシャドーイングに入ると、はるかに効果的です。
- シャドーイングモードを設定する:
- Wait Mode（待機モード）:
+3sまたは+5sを選ぶと、動画が一文を読み終えた後に自動で一時停止し、繰り返す時間が生まれます。完全に手動でコントロールしたい場合はManualを選んでNextを自分で押しましょう。 - Sub Sync（字幕同期）: YouTubeの字幕と音声がずれることがあります。
±100msで調整して、正確なタイミングで追えるようにしてください。
- Wait Mode（待機モード）:
- 声に出してシャドーイングする（最重要）: ここが練習の本質です。文が流れると同時に——または一時停止中に——はっきりと自信を持って声に出して繰り返しましょう。ただ単語を読むだけでなく、話者のリズム、強調、高低、連音をそっくりそのまま真似することが大切です。「影」のように話者に重なるのが理想。Repeat機能を使って同じ文を何度も繰り返し、自然に出てくるまで定着させましょう。
- 徐々に難易度を上げて続ける: 一つのパッセージに慣れたら、さらに挑戦してみましょう。速度を <code>1.25x</code> や <code>1.5x</code> に上げれば、高速の言語反射を鍛えられます。Wait Modeを <code>Off</code> にして連続シャドーイングするのが最も上級で効果的なモードです。毎日15〜30分継続すれば、数週間で目に見える変化を実感できます。