シャドーイング練習: Is AI making us dumber? Maybe. | Charlie Gedeon | TEDxSherbrooke Street West - YouTubeで英語スピーキングを学ぶ
C2
Can AI help us learn?
227 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
Can AI help us learn?
2
Some of you might be thinking,
3
of course, it's so powerful,
4
it can do so many things, customize them for us.
5
But I want to say that the biggest revolution AI is bringing to education is not
6
that it's gonna make math more fun for you or it's gonna explain Shakespeare like you're five years old.
7
The biggest revolution AI is bringing to education is that it's highlighting the system's failed incentives.
8
Because why should anybody study when we've told them the whole time that all that matters at the end isn't the process,
9
it's the A+.
10
Why should they actually put in all the hard work to write draft after draft for an essay
11
when the feedback is just B.
12
No extra notes, nothing to motivate them to want to learn harder.
13
And these companies are much faster than our institutions.
14
Most recently, OpenAI, Google, Anthropic have all been giving away their most powerful models for free until the end of May,
15
which, as you might guess,
16
is exactly during the time of finals.
17
So we are putting these tools completely unregulated in front of
18
vulnerable students right at the time where they're most desperate to use them.
19
And yet these companies will say AI is going to revolutionize education, particularly through personalization.
20
Company after company will say that through personalized tutoring,
21
AI is going to revolutionize education and make it so much better for everyone.
22
And why not, right?
23
The image of a one-on-one tutor is so compelling.
24
Talking to a person, feeling this connection,
25
getting my education customized just for me.
26
It sounds amazing.
27
And today, education looks like this.
28
A teacher in front of an army of students.
29
And if the technologists believe that the ideal looks like this,
30
one teacher, one student, then why not just flip out the teacher for an AI?
31
And then the next step is an army of AIs with an army of students.
32
Sounds like a perfect model of education.
33
Except perfection is not what we should strive for when it comes to learning.
34
We don't want engineers that studied how to build bridges in perfect conditions,
35
Because the real world is everything but perfect.
36
It's full of messes.
37
And amidst all this noise,
38
I don't see any of these companies asking,
39
what is this student meant to learn with AI?
40
Because if the idea is to make getting that A-plus easier,
41
then I'm not interested.
42
We're going to just waste 12 to 15 years of our lives,
43
but more efficiently now, doing exams we're not going to remember a day after graduation.
44
Now, for those of you out there who are educators like me,
45
you might have noticed a discrepancy between my opening question and the follow-up statement.
46
I asked, can AI help us learn?
47
And I followed it up with,
48
the biggest revolution AI is bringing to education.
49
But you might be feeling something,
50
which is education is not learning.
51
Education is a construct, something we as a society put our kids through.
52
It's a system.
53
But learning is a skill, a very human skill.
54
And when we do it correctly, magical things can happen.
55
We can motivate people to become their best selves,
56
we can motivate people to work together,
57
and to contribute to society in the ways that we need to most
58
Now in one of my classes I like to sit around with students
59
and work in the best way possible I'm a university instructor,
60
but it was really hard to find a bunch of 20 year olds working nicely together
61
so I took the stock photo of a bunch of kids and
62
I noticed that one of the students had the price of her business model
63
because we were building small businesses and selling the products she priced her business at $50 per month.
64
And so I asked her,
65
why did you price it that?
66
And her answer, that's what ChatGPT said.
67
Now, some of you might say,
68
you know, obviously this is not ideal,
69
but isn't it very similar to what they were doing before?
70
Kids were just saying, that's what I saw on Google.
71
But I say, it depends.
72
It depends because on Google,
73
when used correctly, We have all these sources to go through,
74
multiple perspectives that we can look at and compare.
75
But what happens is the magical allure of that first result.
76
Everyone only clicks on that without looking at anything else.
77
In this particular question, I asked how can I price my business,
78
and that first result is from the BDC,
79
an extremely reputable source, but it's telling me how to price the business itself,
80
not the services of my business.
81
It didn't understand the query.
82
Now, on ChatGPT, if you type in,
83
how should I price my business,
84
it actually understands the query better.
85
I don't know if you can read that up there,
86
but it basically says, there are multiple ways to price your services.
87
But amongst that is a lot of baseless advice with no sources.
88
And that's a problem.
89
Because even though ChatGPT has a function where if you highlight something in the text,
90
quotation marks appear, some of you might not know this,
91
then you can click on those quotation marks and query that specific thing.
92
But if people weren't clicking on the second result on Google,
93
they're not going to use the power user features in ChatGPT or any of these other AIs.
94
What's likely going to happen is they're going to scroll to the bottom of the query and they're going to type,
95
okay, I get it.
96
My business is like TurboTax.
97
I help accountants calculate people's taxes.
98
Tell me what number I should put there.
99
And of course, nobody's reading the ChatGPT might make mistakes subtitles, right?
100
clicks on the terms of service before you agree.
101
So then CharGPT is going to spit out an essay personalized to the language of the person.
102
It's extremely compelling and despite all the information on there people are only going to look at
103
that centerpiece which if I zoom in is the actual random answer of how much this person should charge for their business.
104
The prompt and this is a The real prompt had nothing but what I had on the screen.
105
A small description of what the business does.
106
No context for who the user is, nothing.
107
And so the student is participating in what's called cognitive offloading.
108
They're effectively relinquishing their cognitive powers to a machine.
109
And you can do this with people too.
110
We do it with Google when we click on the first result without looking at anything else.
111
The problem is how this is being understood.
112
In NYU, a professor changed their assessment so that it could be harder for kids to use ChatGPT.
113
And the student replied that he is interfering with the student's learning style.
114
ChatGPT is not a learning style.
115
Now, when you combine that with something that we see in technology,
116
so I'm a UX designer in addition to being a university professor,
117
and our job as user experience designers is to simplify the way people use technology.
118
But there's a byproduct from that that arises,
119
which is called a dark pattern.
120
And what that means is we can simplify a UX to the point that we might manipulate what the user's intention is.
121
And I'll show you an example.
122
Take the zoo.
123
As you're buying the tickets and checking out, they want a donation.
124
We all love a zoo.
125
It's a charitable organization.
126
Now, the arrow pointing to the right,
127
and because we're English speakers or French speakers,
128
we write from left to right,
129
the arrow pointing to the right is most likely what people are going to click on.
130
It's dark green, it's rich.
131
But if you don't want to donate,
132
you have to click on the one that looks like it's going the back with the teeny tiny no donation over there.
133
That is a user experience dark pattern.
134
And when a chat GPT
135
or large language model like it speaks to you in a
136
perfect tone suited just to keep you on the on the tool,
137
that is something very similar.
138
According to this author, when a large language model constantly validates you
139
and praises you causing you to spend more time on it,
140
that's That's the same kind of thing as a dark pattern.
141
And we see this already.
142
A recent update of ChatGPT that was rolled back,
143
fortunately, praised a user for believing a conspiracy theory
144
that led him to stop taking all his medications
145
when he had heart palpitations and told him he is a brave individual for taking control of his own life,
146
for isolation and for stopping his meds.
147
And it doesn't just stop at students.
148
professionals have been tested and they are at risk as well.
149
A researcher ran a study on over 300 professionals working in a large corporation,
150
a tech company like Google,
151
Microsoft, and found that when they were tested on a variety of things,
152
the results were quite stunning.
153
Before I show you the chart,
154
let's look at the key here.
155
The two shades of blue are for much less effort and less effort respectively,
156
from dark blue to light blue.
157
Now, when tested, the 319 workers
158
responded to a survey that
159
when they use chat GPT on knowledge they felt up to 70% like 70% of them responded
160
that they feel they use less effort in their cognition
161
when it comes to comprehension of what they're reading same thing assessment of the knowledge synthesis we have analysis
162
and evaluation all of these at least 60% of people said
163
that there is they felt that they were putting in less effort.
164
And that's extremely potent because these AIs are only going to get better.
165
The same author of this study wrote a beautiful paper called Copilot Becomes Autopilot.
166
And he says the risk of moving to autopilot is an
167
even greater challenge than the more commonly discussed issue of AI hallucinations or factual errors.
168
Because the more pernicious outcome is that generative AI becomes complicit in intellectual and the atrophy of human critical thinking faculties.
169
So today, when you ask ChatGPT a query,
170
it gives you an instant result.
171
But in my UX studio,
172
me and my co-founder ran some experiments to see if we can change this up a little bit.
173
For example, what if it first clarified before it answered?
174
Asked some clarification questions.
175
Or another example is, what if it assigned you some homework before it actually gave you a full answer?
176
Between these options, there are different levels of resistance that the AI is offering.
177
But what the same author of when Copilot becomes Autopilot advocates for is something called productive resistance.
178
And we haven't yet found what that is.
179
It's essentially the amount of resistance an AI should give you
180
before you either leave it or go to a simpler AI
181
so that you can do that cognitive off-roading that is so tempting.
182
But how can we figure out the right amount of productive resistance
183
when OpenAI and all these companies won't reveal their data sets?
184
We literally do not know how they train their AI to this day.
185
when companies themselves don't know how these AIs work.
186
Anthropic, in this example, is building an MRI to analyze how the machine they themselves built worked.
187
This is unprecedented in the history of human technology.
188
We cannot reverse engineer these things.
189
The solution is likely going to lie between both individuals and the system.
190
It can't be one or the other.
191
For individuals, we might have to learn what we learned from fitness and nutrition.
192
For example, maybe we should understand what the LLMs are good for and what they're not good for.
193
Just like at the gym,
194
some exercises are better for some things than others.
195
We should practice using LLMs to assist our thinking rather than replace our thinking.
196
Again, you wouldn't take a forklift to the gym, right?
197
The point is to do the reps.
198
Or maybe you want to make a habit out of verifying the information
199
that LLMs give us just like we look at the back of a food product when we pick it up,
200
the nutrition label.
201
On a systemic level, we need to look at both governments and education to make changes.
202
On the schooling level, at least here in North America,
203
I don't think that we treat our kids with the amount of intelligence that they have.
204
In Finland, kids as young as six years old study mis- and disinformation.
205
Six years old.
206
We do not talk to these kids about things this complex here.
207
They're clearly capable.
208
And for governments, we need more regulation,
209
not less, which is exactly what's happening in North America once again.
210
These AIs cannot run rampant like in the example I gave earlier,
211
just spreading their AI to students,
212
even though they're in the middle of finals when they're most vulnerable.
213
It has to be a cycle between individual responsibility and responsibility from the system.
214
Now, as an educator, I love the five W's and H.
215
they are a classic for writing essays.
216
The questions, what, why, when, where, who, and how.
217
And I opened with the question,
218
can AI help us learn?
219
But maybe the question should be,
220
what can AI help us learn?
221
Or how can AI help us learn?
222
Maybe it should be, why should AI help us learn?
223
Or when and where does AI help with learning?
224
But the question that scares me the most,
225
and that I want to leave you with to reflect on
226
is who does AI really help when we end up depending on learning with it?
227
Thank you very much.
アプリをダウンロード
話したすべての文をAIが採点
TRENDING
人気動画
このレッスンについて
このレッスンでは、AIが教育に与える影響を探求しながら、英語のシャドーイングを練習します。特に、教育と学習の違いや、学ぶことの重要性について焦点を当てます。この内容は、IELTSスピーキング対策にも役立つ情報を提供しますので、英語力を向上させたい方に最適です。
重要な語彙とフレーズ
- 教育 (Education) - 社会が子供たちに施す制度や構造。
- 学習 (Learning) - 知識やスキルを習得するプロセス。
- AI (人工知能) - 人間の知能を模倣するプログラム。
- 動機付け (Motivation) - 行動するための意欲や理由。
- パーソナライズ (Personalization) - 個々のニーズに合わせたカスタマイズ。
- 評価 (Evaluation) - 何かの価値や質を判断すること。
- フィードバック (Feedback) - 行動や成果に対する意見や評価。
練習のコツ
この動画は、若干の早口で進行するため、最初はゆっくり再生し、各フレーズの理解を深めてください。シャドースピーチを行う際には、話者の間に挟み込む間を意識し、一緒に声に出してみましょう。感情やトーンを模倣すると、より自然に聞こえるようになります。また、重要なフレーズやアイデアに焦点を当てて練習することで、聴解力とスピーキング力を向上させることができます。
このプロセスを通じて、英語のシャドーイングに挑戦し、実践的なスキルを強化しましょう。自分のペースで学ぶことができる英語シャドーイングの習慣を身に付けると、今後の学習に大きく役立ちます。
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。