쉐도잉 연습: What AI can (and can't) currently do in a real lab | Jennifer Doudna - 영상으로 영어 말하기 배우기

연습 모드 선택

각 모드는 전체 화면 연습실로 열립니다.

75 문장

문장이 너무 짧거나 길면 Edit를 눌러 조정하세요.

There's a lot of discussion about AI and biology, a lot, right?

And people are always asking me about this too.

You know, how is AI going to change biology?

How is it going to change the genome editing?

And my immediate answer is, well, it's certainly going to accelerate everything.

It's making our work go a lot faster

because we can increasingly speed up things that used to take a lot more time.

And I think in the future, that'll happen even more.

You know, I do imagine a day

when we can use AI to help us predict which experiments are going to be worth doing.

And that means that we don't have to do as many, we hope, you know.

But I think artificial intelligence absolutely needs human intelligence in biology to be useful.

scientists are increasingly using ai in biology to accelerate the rate of experimentation

and the way they're doing this is in a few ways

they are using it to do very you know kind of mundane types of things

that we used to have to spend a lot of time on

and i'm referring here to things like writing summaries of projects that we need to submit, you know, taking large data sets and culling them,

organizing data into different domains, combining different types of sets of data, converting data from one format into another,

all those types of things used to take huge amounts of time.

Sometimes they still do.

But, you know, increasingly AI is a real assist there.

So that's very interesting.

And lots of people are using it now for that kind of thing.

What I think is potentially more interesting, but it's going to take longer, but we're starting to see this happening,

is that scientists are using AI to make predictions about data,

to discover things in sets of data that might be very hard to discern manually, but where, you know, a machine learning system can go through data

and find trends that would be hard to see by inspecting data yourself,

but a machine system that has the ability to look across large sets of data for small

but significant trends can pick these out.

I think that's very exciting.

And we're seeing this happening more and more.

And I do think in the future, as scientists become better at applying AI to particular questions in biology,

that we're going to see an acceleration of even things like hypothesis generation

and the ability to answer questions that might have taken, you know, a long time previously, but can go fast now because you can use coding agents,

for example, to write code quickly.

So those are where I'm seeing really interesting opportunities even now, and it's going to accelerate in the future.

In experimental biology, it's interesting to think about the role of computation.

I think, you know, when I started out in biology, we used computers to do things that were,

you know, annoying, but, you know, needed to get done and were very hard to do manually.

And I'm thinking about things like converting data from one format to another, combining sets of data together.

You know, I used to be focusing on molecular structure determination using X -ray crystallography, for example.

And that was a field where, at the time, we did lots of,

you know, collecting of X -ray diffraction data from crystals and then combining data from different crystals,

having to convert it into different formats, things of that type.

Even just collating data, you know, things that sound very boring, and they were, but they had to be done and they had to be done right.

So we used computation that way back in those days.

But, you know, increasingly, you know, computers and the models that we're training on data sets are becoming a lot more sophisticated.

So increasingly, we're able to use these for predictions and for the kinds of,

you know, accelerations of experimental science that weren't possible in the past.

Do I think scientists will become, you know, desk jockeys in the future?

No. And the reason is that I don't see robots, you know, this would require really very sophisticated robots.

And I think that robotics right now, even with some of the exciting things that have happened, just are not at the point and frankly,

probably not even close to being capable the way a person is to make changes to experiments on the fly.

We do this all the time in the lab.

It's kind of a natural thing for a graduate student

or a postdoc in the lab to be working on a protocol and realize that, oh, I need to double the volume of this thing

or I need to change the way I'm doing some manipulation of a sample.

That's something that a person can do easily on the fly, whereas for a machine, it would take a lot of effort.

And believe me, I've seen this myself in my lab,

how much work it is to change a robotic format that we have for a robot versus just telling a student, hey, you know, instead of,

you know, manipulating the sample this way, do it that way.

They can change like that.

What I think is going to happen in the future with experimental science is actually

that I think each person who is an experimentalist, like let's say a graduate student,

will effectively have a set of technical staff that are computers or AI agents assisting them.

And those AI agents will be brilliant, but somewhat naive about the actual details of the project.

And so we find this now, you know, that students are using particularly coding agents this way,

where they have coding agents that are set up to help with data manipulation, help them with collecting data in an experiment, even designing how to collect the data

but they don't have the ability to tell the scientist this is the experiment you should be doing

or you should be doing it this way and not that way, again, sort of depending on the details.

But this is really what we're seeing.

And so I imagine that will probably be true for a while, again, because it gets back to this challenge of how do we

ever get enough data to train models to be able to not only kind of regurgitate things that they've been trained on,

but to really be innovative and come up with new ideas.

I don't see that in the near future.

앱 다운로드

유창하게 말하는 데 필요한 모든 것

AI 발음 평가문장마다 점수 확인

IPA 발음 연습모든 소리 마스터

단어 학습나만의 단어장 만들기

단어 게임놀면서 외우기

QR 스캔

Download on theApp Store Get it onGoogle Play

FOR YOU

이 레슨의 어휘와 말하기 포인트

이 C1 수준 말하기 레슨은 영상 “What AI can (and can't) currently do in a real lab”을(를) 바탕으로 합니다. 가장 자주 반복되는 단어는 다음과 같습니다: biology, scientist, increasingly, agents, experiment. 이 영상에는 섀도잉할 문장 75개와 단어 1102개가 있습니다. 말하는 구간의 길이는 6:39입니다. 화자는 분당 약 166단어로, 일상 대화에 가까운 자연스러운 속도로 말합니다. 단어의 84%가 영어에서 가장 많이 쓰이는 3,000단어에 속합니다. 나머지는 연습 전에 미리 확인해 두세요.

이 영상의 핵심 어휘

영상에서 가장 어려운 단어 15개를 발음, 뜻과 함께 정리했습니다.

단어	발음	뜻
acceleration 명사	/əkˌsɛl.əˈɹeɪ.ʃən/	가속
prediction 명사	/pɹɪˈdɪkʃən/	예언, 예측
sophisticated 형용사		정교한
collate 동사	/kəˈleɪt/	함께 합치다
discern 동사	/dɪˈzɝn/	인지하다
postdoc 명사	/ˈpoʊstˌdɒk/	박사후연구원
robotics 명사	/ɹoʊˈbɑ.tɪks/	로봇공학, 로봇학
jockey 명사	/ˈd͡ʒɑki/	기수
manipulate 동사	/məˈnɪp.jʊ.leɪt/	조종하다
genome 명사	/ˈd͡ʒiː.noʊm/	게놈
naive 형용사	/naɪˈiv/	순진하다, 우직하다
hypothesis 명사	/haɪˈpɒθɪsɪs/	가설
organize 동사	/ˈɔɹɡənaɪz/	조직하다
frankly 부사	/fɹæŋkli/	솔직히
predict 동사	/pɹɪˈdɪkt/	예언하다, 예측하다

이 영상의 문법

화자가 가장 많이 쓰는 문형을 영상 속 실제 표현과 함께 정리했습니다.

문형	영상 속 표현
“used to” used to + 동사 — 지금은 아닌 과거의 습관이나 상태	used to take · used to have · used to be
현재완료 have/has + 과거분사 — 과거의 일이 지금도 관련이 있을 때	have happened · I've seen · they've been trained
수동태 be + 과거분사 — 누가 하는지보다 무슨 일이 일어나는지에 초점	be done · are set · they've been trained

주의할 발음

화자는 we're, don't, I'm 같은 축약형과 약화된 형태를 19번 사용합니다. 들리는 대로 짧게 발음하세요.

“sh”와 “zh” 소리: computation /ˌkɑmpjʊˈteɪʃn̩/, acceleration /əkˌsɛl.əˈɹeɪ.ʃən/, manipulation /məˌnɪp.juˈleɪ.ʃən/, prediction /pɹɪˈdɪkʃən/, diffraction /dɪˈfɹækʃən/
긴 단어 — 강세 위치에 주의: accelerate /ɪkˈsɛl.əˌɹeɪt/, manually /ˈmænj(u)əliː/, acceleration /əkˌsɛl.əˈɹeɪ.ʃən/, manipulation /məˌnɪp.juˈleɪ.ʃən/, crystallography /ˌkɹɪstəˈlɑɡɹəfi/

한국어 화자가 어려워하는 소리:

/f/ — ㅍ(/p/)으로 바꾸지 말고 윗니를 아랫입술에 대기: crystallography /ˌkɹɪstəˈlɑɡɹəfi/, diffraction /dɪˈfɹækʃən/, frankly /fɹæŋkli/
/z/ — ㅈ이 아니라 성대를 울리는 /s/: discern /dɪˈzɝn/, organize /ˈɔɹɡənaɪz/
/r/과 /l/ 구별하기: accelerate /ɪkˈsɛl.əˌɹeɪt/, acceleration /əkˌsɛl.əˈɹeɪ.ʃən/, crystallography /ˌkɹɪstəˈlɑɡɹəfi/, frankly /fɹæŋkli/

이 영상으로 연습하는 방법

먼저 말하지 않고 영상을 끝까지 듣고 모르는 단어를 적어 둡니다.
0.75배속으로 한 문장씩 섀도잉을 시작하고, 익숙해지면 보통 속도로 돌아갑니다.
자신의 목소리를 녹음해 원본과 비교하고, acceleration, prediction, sophisticated 같은 단어에 특히 주의합니다.

쉐도잉이란? 영어 실력을 빠르게 키우는 과학적 방법

쉐도잉(Shadowing)은 원래 전문 통역사 훈련을 위해 개발된 언어 학습 기법으로, 다언어 학자인 Dr. Alexander Arguelles에 의해 대중화된 방법입니다. 핵심 원리는 간단하지만 매우 강력합니다: 원어민의 영어를 들으면서 1~2초의 짧은 지연으로 즉시 소리 내어 따라 말하는 것——마치 '그림자(shadow)'처럼 화자를 따라가는 것입니다. 문법 공부나 수동적인 청취와 달리, 쉐도잉은 뇌와 입 근육이 동시에 실시간으로 영어를 처리하고 재현하도록 훈련합니다. 연구에 따르면 이 방법은 발음 정확도, 억양, 리듬, 연음, 청취력, 말하기 유창성을 크게 향상시킵니다. IELTS 스피킹 준비와 자연스러운 영어 소통을 원하는 분들에게 특히 효과적입니다.

섀도잉 방법: 단계별 전체 가이드 읽기 →

쉐도잉 연습: What AI can (and can't) currently do in a real lab | Jennifer Doudna - 영상으로 영어 말하기 배우기

연습 모드 선택

앱 다운로드

추천 영상

AI, Machine Learning, Deep Learning and Generative AI Explained

Very excited to share that I now have a column in New Scientist

Editing Your DNA, with Nobel Prize Winner Dr. Jennifer Doudna

How AI is helping brands identify consumer trends faster

AI & Inequality | Daron Acemoglu: Big Tech poses risks as AI reshapes society

Harnessing generative AI in architecture, engineering & construction

We Can Now Simulate a Human Brain, Scientists Show

When AIs act emotional

AI Will End Philosophy, Physicist Claims

Why the Arab World never united

Alexandr Wang on what matters in an age of abundant intelligence

Be Better than AI: 8 Highest Value Skills to Learn Now