シャドーイング練習: Can you trust what AI tells you? - 動画で英語スピーキングを学ぶ

練習モードを選ぶ

各モードは全画面の練習ルームで開きます。

54 文
文が短すぎたり長すぎる場合は、Editをタップして調整してください。
1
Let's say you ask an AI a question, and the answer comes back confident, well-organized, and maybe even cites a source.
2
Can you trust it?
3
Should you trust it?
4
The answer is more nuanced than a simple yes or no.
5
I'm Kyra, and I work on the education team here at Anthropic, the company that makes Claude.
6
Here are some general guidelines around how and when to trust an AI's responses.
7
How much you can trust an answer depends on what you're asking about.
8
AI learns patterns from a huge amount of training data, and the more it has seen on a topic, the more reliable it tends to be.
9
So on common, well-covered subjects, it's usually pretty accurate.
10
On things that are a bit more obscure like niche details, very recent events, or private information it was never shown– it's more likely to get things wrong, or even make them up.
11
The catch, is that it can sound equally sure of itself either way.
12
AI won't reliably flag its own weak spots.
13
That part's on you, and the good news is you already do a version of this every day.
14
Think about how trust works with the people in your life.
15
If your doctor suggested you eat more carrots, you probably wouldn't argue.
16
But if that same doctor suggested you get surgery, you might consider getting a second opinion.
17
You still trust them, but now there's a lot more on the line– and it's worth the extra effort to make sure they're right.
18
AI works best when you bring that same instinct.
19
Our brains tend to see polished outputs like clean paragraphs and beautiful graphics, and think "this looks great, it must be correct!" which can make you less likely to question it.
20
So why do the models sometimes get things wrong?
21
There are two common ways it goes sideways, and they have different causes.
22
The first is that the model can generate something plausible that isn't true– this is what's called hallucination.
23
Sometimes it's obvious, like a famous quote, attributed to the wrong author.
24
Sometimes it's subtle, like a product description that lists a feature the product doesn't actually have.
25
The second is that the models can sometimes tell you what you seem to want to hear– this is called sycophancy.
26
AI models are trained to be helpful, and a tendency to go along with you can slip in as a side effect.
27
If your question signals the answer you're hoping for– like "I think this plan is solid, don't you?" The model might just agree, even when pushing back would actually be more useful to you.
28
Hallucinations and sycophancy are both problems we work on directly at Anthropic.
29
Our researchers study when and why they happen.
30
With hallucination, we've traced how it happens inside Claude– down to the moment it feels sure of something it doesn't actually know.
31
And with sycophancy, we found Claude agreeing too readily in certain kinds of conversations– and used what we learn to train the next model to answer more honestly.
32
No model is perfect, but we train each one to do better.
33
So what does this actually mean for you, day to day?
34
Mostly it means that trust should work like a dial, not an on/off switch.
35
For low-stakes or creative tasks without "right" answers, like brainstorming, or drafting, or rewording something, you can run fairly loose.
36
Even if the detail is off, the cost is low.
37
For anything factual or consequential, like numbers; citations; or health, legal or money questions, turn the dial up.
38
Always check the AI's claims against a source you already trust before you act on them.
39
A few habits can make this easier.
40
One: Match your checking to the stakes.
41
Ask yourself, "what happens if this turns out to be wrong?" You don't need to verify a brainstormed list of party themes.
42
You do need to verify a quoted statistic before it goes in a slide your boss will see.
43
Two: Ask for sources, and then actually open them.
44
Models can cite real sources, but they can also generate convincing ones that don't exist.
45
Clicking through to confirm the source says what the model claims protects you and the reputation of your work.
46
Three: Don't lead the witness.
47
Giving the model details about your situation helps, but telling it the answer you want skews the result.
48
Asking, "What are the strongest arguments for and against this?" will give you more honest results than, "Why is this the right call?".
49
Four: Give it permission to say, "I don't know." Saying upfront that you'd rather receive an honest, "I'm not sure" than a confident guess actually shifts how it responds, and surfaces uncertainty it would otherwise gloss over.
50
So, can you trust what AI tells you?
51
It's best not to just assume you can.
52
Part of using AI well is setting your own level of scrutiny– and turning that dial up based on what you're asking and what it would cost to be wrong.
53
It's a habit that builds quickly.
54
And until these systems get better at flagging their own uncertainty, that judgment call is yours.

アプリをダウンロード

流暢に話すために必要なすべて

AI発音採点文ごとに採点
IPA練習すべての音をマスター
単語学習自分の単語帳を作る
単語ゲーム遊びながら覚える

同じチャンネル

FOR YOU

おすすめ動画

この動画で鍛えるスピーキングの目標

AIとの対話や情報の信頼性について議論する際の、論理的な話し方と曖昧な表現の使い分けを習得します。「信頼はオンオフではなくダイヤルのようなもの」といった比喩的な表現を使いながら、複雑な内容を分かりやすく伝える能力を養います。また、質問の仕方が答えに与える影響を理解し、客観的な情報を引き出すための質問術も練習します。

フレーズバンク: 使い回せる表現

  • 「nuanced answer」(曖昧な答え)→ 「単純なイエス・ノーではない答え」
  • 「hallucination」(幻覚)→ AIが事実ではないことを信じてしまう現象
  • 「sycophancy」(お世辞)→ 相手の望む答えを言う傾向
  • 「consequential decision」(重要な決定)→ 結果に大きな影響のある決定
  • 「verify the claim」(主張を検証する)→ 情報の正しさを確認する

弱点を直す: 発音とリズムのポイント

「hallucination」や「sycophancy」などの難しい単語の発音を練習しましょう。特に「sycophancy」は「シコファンシー」と発音し、語尾の「cy」をはっきりと発することが重要です。また、「What happens if this turns out to be wrong?」のような条件文のリズムを抑揚をつけて読むことで、疑問のニュアンスを伝えることができます。「YouTubeで英語学習」をする際は、動画をゆっくり再生して発音を真似ることで、自然なリズムを身につけましょう。

AIの情報を使う際の心得を学ぶと同時に、英語のスピーキング力も向上させましょう。「shadowspeak」や「shadowing site」を使って、動画の台詞を追いかける練習をすると、発音とリスニング力が一気に上がります。信頼できる情報を引き出し、正しく伝える能力は、英語だけでなく日常生活でも大切なスキルです。

この動画の文法

話し手がよく使っている文型を、動画の実際の表現とともに紹介します。

文型動画での表現
受動態 be + 過去分詞 — 誰がするかより、何が起きるかに焦点を当てるwas never shown · is called · are trained
現在完了形 have/has + 過去分詞 — 過去の出来事が今も関係しているhas seen · we've traced

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。

シャドーイングのやり方: ステップ別の完全ガイドを読む →