00:00:04
Your grandchildren will be
the last generation to read and write.
00:00:10
I know that sounds strange.
00:00:12
Almost unthinkable.
00:00:14
Text is everywhere around us.
00:00:16
We use it hundreds of times
every single day
00:00:19
and it's woven into the fabric
of our daily lives.
00:00:23
But today, I'm going to make the case
00:00:26
that humanity's relentless pursuit
of better ways to convey ideas
00:00:31
and preserve knowledge
00:00:32
doesn't end with text.
00:00:35
I think we're at the dawn of a new era
of AI-enabled communication,
00:00:40
and I think that future generations
will slowly replace text
00:00:44
with more intuitive forms
of communication,
00:00:46
like audio, video,
00:00:48
and eventually immersive technologies.
00:00:50
And one day, I think we'll look back
at reading and writing
00:00:54
as historical artifacts,
00:00:55
like we do with papyrus scrolls
or hieroglyphs, or cave paintings.
00:01:01
Don't get me wrong, I love reading.
00:01:03
This is not a personal
vendetta against text.
00:01:06
(Laughter)
00:01:07
Some of my fondest memories
are walking around in my local library,
00:01:10
back in Copenhagen
like you see behind me here,
00:01:13
picking out all kinds of books,
00:01:15
mostly science fiction books,
00:01:16
returning home to read them
00:01:18
so I could go back there again
00:01:20
and pick out the next set
of exciting books to read.
00:01:23
I loved disappearing into these worlds
00:01:25
that were different
than the physical world around me.
00:01:27
And that, of course, only was amplified
00:01:29
when I discovered the internet
at around ten years old,
00:01:32
which opened up
an entirely new world of ideas,
00:01:36
of music and people.
00:01:38
On the internet, information was free,
00:01:40
and I saw firsthand
00:01:42
how technology didn't just change
the distribution of content,
00:01:45
it changed the content in itself.
00:01:48
The difference between a lively forum
00:01:50
and a book
00:01:52
is massive.
00:01:54
A blog and a newspaper, and so on.
00:01:57
In music, which is my
big passion outside of work,
00:02:01
I saw how software instruments
and sampling and drum machines
00:02:05
gave birth to entirely new genres
that had never been possible before.
00:02:09
And not only that,
but all the people around the world
00:02:12
making new, exciting music
00:02:14
could share it with the world
without the middlemen of labels.
00:02:18
The video games I played
got connected to the internet.
00:02:21
You formed communities with people
around the world that you didn't know.
00:02:25
That's what my parents thought, at least.
00:02:28
I started my first business
when I was 13 years old
00:02:31
in World of Warcraft.
00:02:32
We were 50 people.
Slaying dragons together online.
00:02:36
And I would later learn in life
that it's actually not that different
00:02:39
than running a startup.
00:02:40
(Laughter)
00:02:42
And those early years of my life
00:02:43
sparked a lifelong interest
in media and technology.
00:02:47
How they changed the way that we create,
consume, play and communicate.
00:02:52
And in 2016, I discovered a research paper
00:02:55
called "Face to Face" by Professor
Matthias Niessner and his team.
00:02:59
They built a system that,
using neural networks,
00:03:02
could produce really, really
photorealistic video.
00:03:05
And when I saw this for the first time,
I felt like I'd seen magic,
00:03:08
and I was convinced that in ten years,
00:03:10
you're going to be able
to create a Hollywood film
00:03:13
from your bedroom without needing
anything else than just your imagination.
00:03:16
That took three years,
00:03:18
and I think that's actually
going to hold up.
00:03:20
I couldn't get this idea out of my mind,
00:03:23
and eventually I ended up
founding Synthesia,
00:03:26
an AI video company,
00:03:28
along with Matthias and Lourdes
and Steffen, my cofounders.
00:03:32
And this really was a way to kind of marry
my interest in media and technology
00:03:37
and sci-fi.
00:03:38
We started the company with the vision
00:03:40
of making everyone in the world
into a Hollywood director.
00:03:43
And while that's definitely
still a very exciting vision,
00:03:46
as the years kind of went on,
00:03:47
we realized that as exciting
as AI-generated Hollywood films are,
00:03:51
it's only the tip of the iceberg.
00:03:53
The really exciting potential
about these technologies
00:03:56
is that they're going to enable
every single piece of content,
00:03:59
from text messages to novels
to boring corporate training materials,
00:04:03
to be brought alive in video and audio.
00:04:06
But before we talk more about video,
let's talk about text.
00:04:10
Text is the original way of compression
for human communication.
00:04:14
We encode our thoughts and knowledge
00:04:16
into symbols that can convey meaning
across time and space.
00:04:20
That hasn't always been the case.
00:04:21
Thousands of years ago,
text didn’t exist,
00:04:23
and the only way we had
to share information was orally.
00:04:27
In 1500 before Christ,
the first alphabet was invented.
00:04:30
And it essentially took
these very complex writing systems
00:04:33
and simplified them
into just a few characters,
00:04:36
laying the groundwork
for the modern society that we know today.
00:04:40
In 1440, Gutenberg
invented the printing press,
00:04:43
and for the first time in history,
we could mass produce written content.
00:04:48
It would take until the mid-20th century
00:04:51
before reading became
something for the masses,
00:04:53
and people slowly became literate.
00:04:55
If we fast-forward to today,
text is ubiquitous,
00:04:59
and it's impossible to live a life
without being able to read and write.
00:05:03
But as great as text is,
00:05:05
it's an imperfect technology.
00:05:07
It's very efficient
and it's very scalable,
00:05:10
but it's a very, very lossy method
of compressing information.
00:05:14
It lacks all the nuance
and additional information
00:05:16
that we get when we speak
to someone in real life.
00:05:19
Your tone of voice,
your body language, where you are.
00:05:23
All those things matter for that message.
00:05:26
Text can be interpreted
a million different ways
00:05:28
depending on the receiver.
00:05:30
Is this good news or bad news?
00:05:33
It's very hard to tell, right?
00:05:34
(Laughter)
00:05:36
But we invented emojis
to make it a little bit better.
00:05:38
(Laughter)
00:05:43
But even emojis are not perfect.
00:05:45
People are still debating
if this smiley is happy.
00:05:50
Is it pissed?
00:05:51
(Laughter)
00:05:53
Is it planning to kill you?
00:05:54
(Laughter)
00:05:56
The debate goes on.
00:05:59
And if you think
about visual communication,
00:06:02
this is an image.
00:06:03
It took all of you a few seconds
00:06:05
to look at this image
and understand it, right?
00:06:08
It's a very intuitive way
of consuming information.
00:06:12
What if I share this with you instead?
00:06:13
A text description of the image.
00:06:15
Would have taken you
30 seconds to read it,
00:06:18
with high cognitive load,
00:06:19
to take pieces of those symbols
and turn them into image an in your mind.
00:06:24
And even if I gave you
30 pages of description,
00:06:27
you would always imagine a different image
00:06:29
than what you're seeing
behind me right now.
00:06:31
Once we add a time dimension, as in video,
00:06:34
this problem massively compounds.
00:06:37
So it's not that strange
that since the invention of text,
00:06:40
we've been innovating
towards richer and more intuitive ways
00:06:43
of exchanging information.
00:06:44
We invented radio, TV,
internet, VR, social media,
00:06:48
and now, AI.
00:06:50
In 2024, it's very obvious
that people want to watch and listen.
00:06:54
If you look at the digital economy,
00:06:56
TikTok is the fastest growing
social network.
00:06:58
It's also the fastest growing
search engine,
00:07:00
which is very interesting.
00:07:02
Video and audio is everywhere
in the apps that we use.
00:07:05
We send voice notes on WhatsApp.
00:07:07
Now we're on dating apps.
00:07:08
When we shop, we watch product videos.
00:07:11
And my thesis is
that the more we consume video,
00:07:15
the more bored we are by text.
00:07:17
And I have to say,
this is also true for me.
00:07:20
Even though I love reading,
00:07:21
when I long to learn something new,
00:07:23
I usually start on YouTube
or on TikTok, or listen to a podcast.
00:07:28
And only if I'm really, really
invested in something,
00:07:30
I'll take the hours out of my day
to read a 200-page book.
00:07:34
It just doesn't feel that much
like it's worth it anymore.
00:07:36
And for a lot of you,
you probably feel the same.
00:07:39
Do you want to learn music theory
from a long book
00:07:42
or from a video on YouTube that has audio?
00:07:45
Do you want to listen to the news
on a podcast on the way to work,
00:07:49
or fold out this physical
piece of paper somewhere?
00:07:52
Most people feel like this,
00:07:54
but we all have this guilt.
00:07:56
I have at least.
00:07:57
I feel guilty when I watch videos
and I listen to podcasts
00:08:01
instead of picking up
a good old-fashioned book.
00:08:03
And you hear the commentary on this.
00:08:05
"Young people are unable to focus anymore.
00:08:08
They need constant dopamine hits
from cheap content
00:08:11
that they scroll through
on their social media apps.
00:08:14
They don't get outside
their room anymore."
00:08:16
Exactly the same thing
parents have been saying for 200 years.
00:08:20
I have a provocative idea.
00:08:23
What if we're all just tired
of overly dense,
00:08:27
slow information?
00:08:29
Books with too many pages.
00:08:31
Newspaper articles with filler.
00:08:33
What if we've become much more sensitive
to the quality and the conciseness
00:08:37
of the content that we consume,
00:08:38
because we now have infinite choice?
00:08:41
What if the current generation of kids
00:08:43
are able to learn
and absorb information much faster
00:08:46
because of technology, not despite it?
00:08:49
Is the problem us or is the problem text?
00:08:53
We still read a lot, right?
00:08:54
We just read from many
different sources every single day.
00:08:57
Not just a book and a newspaper.
00:08:58
We read our messaging apps,
our social media networks and our blogs.
00:09:02
We listen to long form podcasts.
00:09:04
That's very growing, right?
00:09:06
But we still have this idea that books
are morally superior in some sense.
00:09:10
And even though I'm giving you this talk,
I still very much feel it.
00:09:13
I don't know why that is.
00:09:15
I'll let someone else
do the research on that.
00:09:17
But it's definitely very interesting
00:09:19
how we psychologically attribute
more value to the written word.
00:09:22
So why is so much information still text
if we really prefer video?
00:09:26
It's pretty simple. The answer is cost.
00:09:29
Cost in time and cost in money.
00:09:32
Today we have to pick between
the speed and scale of text
00:09:36
or the accuracy and engagement
of video content.
00:09:39
And so there is this underlying
economic incentive
00:09:42
where only the content and the ideas
and the knowledge we deem important enough
00:09:45
gets converted into video
and audio formats.
00:09:49
And so in the entertainment industry,
00:09:50
this means that we get
"Fast and Furious 278"
00:09:53
rather than avant-garde cinema
from film students.
00:09:57
In the corporate world,
00:09:58
this means that
the Super Bowl ad is a video,
00:10:00
but the fire safety video
is a long, boring document.
00:10:05
And this is about to change
00:10:09
in a big, big way.
00:10:11
AI will change that equation completely.
00:10:14
With AI, we can get both speed,
scale, accuracy and engagement.
00:10:18
AI can create highly photorealistic
content digitally.
00:10:21
Computers can learn
what the world looks like,
00:10:24
and they can replicate it
and remix it in amazing detail.
00:10:27
This is going to usher in
a new wave of creativity,
00:10:29
and it's not going
to be driven by Hollywood.
00:10:31
It's going to be driven by YouTubers
and young people with great ideas
00:10:35
who will take these tools
and tell amazing stories.
00:10:39
At Synthesia we focus on AI avatars,
00:10:42
digital humans
that sound and look like us.
00:10:44
They can even be us
if you make a clone of yourself.
00:10:48
Today, our avatars already interact
with millions of people every single day.
00:10:52
They teach school subjects,
they onboard restaurant workers,
00:10:55
provide health guidance and sell products
in more than 130 different languages.
00:11:00
And they're getting really good.
00:11:01
Very soon they'll be very difficult
to distinguish from reality.
00:11:05
(Video) Hello, I’m Alex,
a personal avatar.
00:11:08
I am so excited and happy to be here.
00:11:10
Can you see how expressive I look
and hear how natural I sound?
00:11:14
Isn’t it great? Bye for now.
Hope to see you soon.
00:11:17
Victor Riparbelli: This video is
entirely AI-generated.
00:11:20
(Video) I'm Alex --
00:11:21
(Laughter)
00:11:23
VR: They still don’t know
when to stop speaking.
00:11:25
(Laughter)
00:11:26
So with these technologies,
00:11:27
we can create anything
without the need for cameras.
00:11:30
We can bring our imaginations to life
00:11:32
without the traditional barriers
of skill and cost.
00:11:35
But what's more exciting
about all these new technologies
00:11:38
is that they're going
to make it extremely easy
00:11:40
for anyone to create content.
00:11:43
We've already seen
how keyboards and computers
00:11:45
turned all of us into writers.
00:11:47
We've seen how PowerPoint
turned all of us into designers.
00:11:50
And with AI, everyone
is going to be able to be a director,
00:11:53
producing Hollywood-grade video
without needing any training at all.
00:11:56
Like you can see here behind me.
00:11:58
At Synthesia, we've already turned more
than a million people into video creators,
00:12:02
and it's really fun to see all the things
people make videos about,
00:12:06
but they don't have to ask their boss
because they can just log on
00:12:09
and make the videos themselves.
00:12:11
Why do people make videos?
00:12:13
It's not a very well-researched area yet,
00:12:15
but we've been working with UCL in London
00:12:18
to figure out how do we learn differently
with video and AI video versus text?
00:12:24
We did a study with 400 participants.
00:12:26
It's still ongoing.
00:12:28
We plan to publish the results
sometime next year,
00:12:30
but there's some really
interesting initial findings.
00:12:33
When you just ask people,
00:12:34
do you prefer to learn
with AI video or text,
00:12:36
the results are pretty astonishing.
00:12:38
77% of people prefer
to learn through text --
00:12:41
sorry, through video.
00:12:43
(Laughter)
00:12:45
So faster and easier content creation
will be transformative.
00:12:49
But with new technologies,
we always invent new media formats.
00:12:54
And right now, most AI-generated content
is what I call a bridge genre.
00:12:59
We're using it to create old formats
with new technologies.
00:13:03
Just like the first newspapers --
00:13:05
sorry, the first websites
looked like newspapers on a screen,
00:13:08
AI videos today are linear,
00:13:11
they have a beginning and an end,
00:13:12
and they essentially emulate
what we can record with a physical camera.
00:13:18
Once we combine AI video
00:13:20
with reasoning systems,
like language models,
00:13:23
we're going to unlock
an entirely new type of media
00:13:25
that's going to be interactive
and personalized.
00:13:28
It's going to be able to think and narrate
and personalize content for us.
00:13:32
If you're learning music theory,
00:13:33
you will have an assistant
that knows your skill level,
00:13:37
knows your taste in music,
00:13:38
and builds a curriculum around that.
00:13:40
All of your kids maybe will have
their favorite celebrities
00:13:43
teaching them math in school.
00:13:47
And they'll do it in a context
that's interesting for your kid.
00:13:50
Maybe that's soccer,
or sci-fi, or whatever.
00:13:53
Education is going to be turbocharged
with these new AI systems.
00:13:58
Entertainment is going to change.
00:14:00
It's going to be shaped by the viewers
and by the world around them.
00:14:04
We'll probably have interactive films,
we'll have TV series that never end.
00:14:09
It sounds weird, but in many ways,
00:14:11
this is what a social media
feed is, right?
00:14:13
It's just an endless story.
00:14:15
Your TikTok feed is also an endless stream
of personalized content just for you.
00:14:21
And when we combine this
with new form factors like AR, VR,
00:14:25
brain-computer interfaces,
00:14:27
media is increasingly going
to appear like real life.
00:14:30
Maybe we'll be transmitting
messages to our friends
00:14:33
just via thought in the future,
00:14:35
and our inbox is not going to be text,
00:14:37
it's going to be videos
00:14:39
of all your friends' and your relatives'
AI avatars talking to you.
00:14:42
It sounds weird,
but I don't think it's that far off.
00:14:46
And I promise you
that your fire safety video at work
00:14:49
will be a Spielberg-esque masterpiece
in the not so distant future.
00:14:52
(Laughter)
00:14:53
These technologies open up
a Pandora's box of questions.
00:14:57
There's ethical questions,
political questions,
00:15:00
design questions,
and there's commercial questions.
00:15:03
And there are many speakers today
sharing great viewpoints in those.
00:15:07
I'll leave you with a few
provocative questions.
00:15:11
Do you care if our content
is AI-generated or not?
00:15:15
Do you care when a film uses
special effects or green screen?
00:15:20
Do we care that actors in commercials
are paid to play someone else
00:15:25
and endorse products that they've
probably never heard of?
00:15:29
Do we care if actors
are real or generated?
00:15:33
Are you a fan of Harry Potter
or Daniel Radcliffe?
00:15:37
Does it matter that Daniel Radcliffe
exists in the real world?
00:15:40
(Laughter)
00:15:41
You'll never meet him. Probably, right?
00:15:45
What about Mickey Mouse
or Pixar characters?
00:15:48
Can we feel something for them
despite them being AI-generated?
00:15:51
Will we trust AI agents,
00:15:52
or do you still want to call
the call center to speak to a human?
00:15:56
Can we be entertained by computers
interacting with each other?
00:16:00
Would you want to play computers,
00:16:01
play soccer, or play chess
against each other?
00:16:04
Maybe not.
00:16:06
The future is ahead of us.
00:16:07
And while a lot of this
sounds like sci-fi,
00:16:09
it really isn't that far off.
00:16:11
And it's up to all of us
in this room and society
00:16:14
to make sure that we build
a really awesome future.
00:16:17
I think we have the technologies
00:16:19
and I think we have the people to do it.
00:16:21
Thank you.
00:16:22
(Applause)
