Home
Đăng nhập
Đăng ký
Luyện nghe
Luyện nghe
/
Video
/
TED Talk
/
How AI Is Breaking the Internet (and What to Do About It) | Matthew Prince | TED
How AI Is Breaking the Internet (and What to Do About It) | Matthew Prince | TED
Chọn chế độ học:
Xem phụ đề
Chọn từ
Viết lại từ
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Phụ đề (296)
0:04
The internet is an absolute miracle.
0:07
And for the last 30 years,
0:09
it has been driven by [a] very consistent
0:13
and fundamental business model
0:15
that has given us everything amazing about the internet
0:18
and everything not so amazing as well.
0:21
I run a company that provides a lot of the internet's infrastructure
0:25
and as a result, means three things.
0:27
One, we've had this unique view into what is going on online.
0:34
Two, we care enormously about the success of the internet.
0:39
And three, I can share with you
0:42
some of how that underlying business model is changing
0:45
and why it's important.
0:47
Because as that business model changes,
0:49
it's going to have a dramatic impact on what the internet is,
0:53
but beyond that, on what content is created,
0:56
on what journalists get compensated for,
0:58
on how researchers are able to do their research.
1:01
And so it's up to all of us
1:02
as that business model of the internet changes,
1:05
to make sure it changes for something we want
1:09
and can learn from the mistakes that we had in the past.
1:12
So to review the past, let's look at what that business model was.
1:15
And we all know it basically, which was: the internet was built on
1:19
we took an idea, and we put it online,
1:21
and then we drove traffic to that idea.
1:25
And then from that traffic, we either sold things,
1:28
subscriptions or ads.
1:31
And it's been pretty amazing to watch
1:34
how that has very much driven the growth of the web
1:38
and the growth of the internet over time.
1:40
And it survived a number of dramatic platform shifts.
1:43
We all started browsing the internet on browsers,
1:45
on our desktop or laptop.
1:46
That transitioned to be increasingly social media,
1:49
which was the interface through which we consumed content.
1:53
And then we went to consuming it through our mobile devices.
1:56
And yet, through all of this,
1:58
the humble banner ad has basically financed it all.
2:01
Traffic has been the currency of the internet,
2:04
and that's what everyone has built their content to try to achieve.
2:09
And if we want to look for who's the Superman,
2:12
who's the superhero that has driven this
2:15
for the last 27 years,
2:17
that has been Google.
2:19
Google built not only the search engine to index how the internet works,
2:23
but then they built the monetization system
2:26
to actually ensure that content creators got paid
2:31
for the traffic that they generated.
2:33
The internet owes a ton to Google in the last 27 years.
2:37
But the world is changing.
2:39
AI is the topic du jour,
2:41
and what I think a lot of people don't understand
2:43
is that AI is yet another platform shift.
2:46
Think about how much more of the way that you consume information already
2:50
is looking at your AI chatbot,
2:52
and how much more of that will be the case in the future.
2:55
We can see it in a lot of things we do kind of normally,
2:58
like, this is a Google search from the early days of Google.
3:02
Not even that long ago,
3:03
where I would type something in, like, when was the first TED conference?
3:07
And then I would get back basically a treasure map
3:10
where I would have to continue to do the work
3:12
and click on the links, generating traffic
3:14
to find what my answer was.
3:16
The search wasn't over when the results returned.
3:19
That was just the beginning, and you had to follow the treasure map.
3:22
That's what a search engine is.
3:24
Today we have something very different.
3:26
We have something that's closer to what I would call an answer engine,
3:29
where you're getting the answer directly on the page.
3:32
And what's amazing about that is that as a consumer,
3:34
it's better for you.
3:36
But what's terrifying about it
3:38
is that it means traffic isn't going to the publishers' sites,
3:42
and traffic is the currency of the internet.
3:45
And so 18 months ago, publishers called us and said, “Are you seeing this, too?”
3:49
And I have to confess,
3:50
we weren't really paying a ton of attention,
3:52
but we went into our logs, we went into our data,
3:55
we pulled out that information and said,
3:57
“How much harder is it to get traffic
3:59
based on the same amount of content today
4:02
than it was with the Google of old?”
4:04
And the answer is, it's 5.2 times harder
4:07
to get traffic for the same amount of content that was out there.
4:10
But that's the good news.
4:11
That's with Google.
4:12
What is it with OpenAI?
4:14
750 times harder.
4:16
What is it with Anthropic?
4:17
32,000 times harder.
4:20
The answer engines are giving the answers
4:22
and people aren't following the footnotes.
4:24
And this was 18 months ago.
4:26
What’s happened since, has it gotten better or worse?
4:28
Do you trust AI more today than you did 18 months ago?
4:32
The answer is most people do, and as a result,
4:36
it's actually gotten twice as hard to get traffic to content that's there.
4:40
You're not following the footnotes anymore.
4:42
You're actually trusting that AI systems are giving you the right answer,
4:45
and that's destroying media.
4:46
Across the board, we're seeing reported numbers
4:49
... of between a 90 and 30 percent decline
4:52
in the traffic to their sites,
4:54
and this is extremely problematic
4:56
because I think journalists deserve to be able to eat.
4:59
(Applause)
5:01
And fundamentally, if the internet is built on traffic
5:05
and this new business model isn't delivering traffic,
5:09
then I'm not sure how media is going to survive
5:12
unless we create a new business model.
5:14
And the good news is the AI companies want journalism.
5:18
They want people out there doing research.
5:20
We all want journalism.
5:21
We want people out there doing research.
5:23
And so I think, I'm actually kind of optimistic
5:26
that there might be a solution to this problem.
5:29
And here's the way I think about it.
5:30
For the first time in human history,
5:32
we have a mathematical model of all of human knowledge.
5:35
It's not perfect, but that's essentially what all of the LLMs are.
5:39
And I think of it like a giant block of Swiss cheese.
5:41
There's a lot of cheese, that's what we know.
5:43
But we also have a lot of holes in the cheese.
5:46
Those are the things we don't know.
5:47
And the thing that is incredible
5:49
is that what all of you want, what I want,
5:51
what the AI companies want
5:53
is they want people out there filling in the holes in the cheese.
5:58
They don’t want you writing stories that already exist --
6:01
they want you actually creating new knowledge.
6:03
And the AI companies, I've talked to every one of them,
6:05
they're all willing to actually pay for people who are creating new knowledge,
6:09
creating new ideas, ideas worth spreading.
6:12
That's what this new system would reward.
6:14
Compare that to what the business model of media is for the last 10 years,
6:20
which is not filling in the holes in the cheese,
6:22
but instead, basically, lighting the cheese on fire, right?
6:27
Incendiary cheese, right?
6:29
Yet another story about what happened at 1600 Pennsylvania Avenue,
6:33
told in the most inflammatory way possible.
6:35
I've sat with media executives
6:36
who brag about the fact that they create content for as little as possible,
6:41
and then write headlines that create the largest cortisol response
6:45
so that you will click on them.
6:47
That's not a world we should be celebrating.
6:49
That's not a world that's making the world better.
6:51
If, on the other hand, we actually celebrated
6:54
who is making the cheese bigger,
6:55
who's filling in that --
6:57
Imagine in the future that you actually had AI companies
7:00
that were measuring
7:01
who actually advanced human knowledge mathematically
7:04
over the course of the last year the most,
7:06
and we gave them an award.
7:08
We could call it the Big Cheese award.
7:09
(Laughter)
7:11
I'm not kidding.
7:12
Like, how amazing would that be?
7:13
And you might be thinking, OK, Matthew, you're crazy.
7:16
Like, I'm the person who thinks AI is just destroying the world,
7:19
and technology always makes things worse.
7:21
I'm going to share just two anecdotes that make me have some hope
7:24
that this might actually be the case.
7:26
The first starts 23 years ago.
7:28
23 years ago, the music industry had a huge problem:
7:30
Napster, Kazaa -- everyone was stealing music.
7:33
And then Steve Jobs stepped on stage, almost 23 years ago to this day,
7:37
and he said, “We’re going to do something different.
7:39
We're going to charge 99 cents per song.
7:41
We're going to actually say that you should be paying
7:44
for the music that you consume.”
7:45
Now that’s not the business model that won,
7:47
but it was the start, and it turned into what is today Spotify
7:51
and all the other systems that are out there,
7:53
which is basically $10 a month, all-you-can-eat music,
7:56
and it's pretty amazing.
7:57
Last year, Spotify alone sent over $10 billion to music creators
8:03
in revenue.
8:05
And again, that's in addition to titles, TikToks, YouTube,
8:09
Apple Music, everybody else,
8:12
more money is going to music creation than any time in human history.
8:15
So technology can be pie-expanding.
8:18
In fact, I thought it was so important
8:20
that I flew up to Stockholm to meet with Daniel Ek,
8:23
the founder of Spotify, and he said, "You know, it's pretty amazing.
8:26
We know when someone goes to Spotify and they type in Taylor Swift,
8:30
"Shake It Off"
8:31
that we can give them what they're looking for."
8:33
But every once in a while, someone comes in
8:35
and types in something like,
8:37
'I want a song about how much fun it is to dance disco with your cat.'
8:40
(Laughter)
8:41
And nobody's written that.
8:44
But they know that nobody's written that.
8:46
And so they actually go back and take those queries
8:48
and then publish them back to music creators.
8:50
And this is the moment you're all going to think you're doing life wrong.
8:54
But there's a guy who lives in Denmark
8:56
who makes €40 million a year
8:57
writing songs for unfulfilled Spotify queries.
9:00
(Laughter)
9:02
And I think that's beautiful.
9:04
That's filling the holes in the cheese.
9:06
Because when I look for a song,
9:07
what I'm looking for is an emotion.
9:09
And this person is helping me find a song to resonate that way.
9:12
And I think that's a model for what could happen as well.
9:15
Another story starts a little bit closer to home.
9:17
So this is the Park Record newspaper.
9:22
The first place I ever saw my name in print --
9:25
my hometown newspaper from Park City, Utah.
9:27
It was dying.
9:28
So three years ago, my wife and I bought it.
9:31
We did not buy it because we thought it was going to be a great business.
9:35
We bought it because we thought local news was important.
9:37
In fact, we had no financial KPIs --
9:39
the only KPI we focused on was how much could we increase voter turnout.
9:43
And the answer was, we got it from the low 20 percent to over 60 percent
9:46
in the last election in three years,
9:48
which is amazing, right?
9:49
(Applause)
9:51
Local news matters.
9:52
But here's the crazy part.
9:55
Terrible business, right?
9:57
Except it's not.
9:58
We're likely going to make more in 2026
10:02
off licensing our content to AI companies
10:04
than we do off digital advertising. Why?
10:06
Because if you're traveling to Park City, Utah,
10:08
and you want to know what's the great new restaurant,
10:11
what's the cool band that's coming to town,
10:13
we're the only ones that have it.
10:15
So think about how amazing that would be.
10:17
Think about how if in the future,
10:18
the business model of the new AI-driven internet
10:21
actually rewards the media we're all craving,
10:24
local, unique, original content.
10:27
That's what I'm super excited about.
10:29
But there are problems.
10:31
Challenges.
10:32
It's kind of like in the superhero films,
10:35
where the hero of the last movie becomes the villain of the next.
10:39
And so our good friends at Google ...
10:42
Ugh.
10:43
(Laughter)
10:44
They are holding this back.
10:46
They believe that they have a God-given right
10:49
to all of the content online.
10:50
They shouldn't have to pay for any of it.
10:52
And they have a huge advantage.
10:54
In fact, it's hard to believe this,
10:55
but every other AI company was actually started
10:58
to oppose Google, right?
10:59
That’s what OpenAI started with.
11:01
Anthropic was just a spinoff of OpenAI, xAI.
11:03
All of this was all to say, we can't let Google win this entire game.
11:07
How much of an advantage do they have?
11:08
Well we looked at this again,
11:10
and we found that every single page that OpenAI sees,
11:13
Google sees three.
11:15
Four in Microsoft’s case,
11:18
five in Anthropic's case,
11:19
21 or 22 in xAI's case.
11:22
An enormous advantage that Google has,
11:25
and they want to hold on to that advantage.
11:27
In fact, they're playing incredibly dirty.
11:29
They're saying to content creators,
11:31
"You can't stop us from training on your content
11:34
or we will take you out of search."
11:37
Horrible.
11:38
Superman, how far you’ve fallen!
11:40
(Laughter)
11:42
And so I was pretty depressed, I have to confess,
11:45
thinking we were on the cusp of creating something amazing:
11:48
a better internet.
11:51
And yet, I have to just realize that no AI company's top priority
11:56
is paying content creators.
11:58
What's their top priority?
12:01
Beating Google.
12:03
Which gave me an idea.
12:05
Kind of a wonderful idea.
12:08
An idea which is going on right now, and it's working.
12:10
Which is:
12:12
don’t charge for all the content that’s out there,
12:14
just charge the AI companies for the ability to catch Google.
12:19
And I don't think that's the end game,
12:21
but I think it might be the iTunes moment.
12:23
It might be the point that shows us content is worth paying for.
12:27
Everyone should be out there,
12:28
and we can create a better business model for the internet.
12:31
And so I'm hopeful that on this stage, not too long from now,
12:35
I get to hand out one of the first Big Cheese awards,
12:37
and I really appreciate it.
12:38
Thank you.
