The Incredible Creativity of Deepfakes — and the Worrying Future of AI | Tom Graham | TED

Summary

In a TED Talk, Tom Graham discusses the capabilities of his company, Metaphysic, which specializes in creating hyper-realistic artificial content using AI. He demonstrates deepfake technology by transforming himself into Tom Cruise and others in real-time, raising concerns about authenticity in media. Graham also explores the potential benefits, such as enhanced emotional connections and educational applications, while emphasizing the need for individuals to own and control their data to prevent misuse.

Takeaways

• 😀 The company became famous for a deepfake video of Tom Cruise that gained massive online views.
• 🎥 The speaker is from Metaphysic, a company specializing in creating AI-generated content that appears realistic.
• 🤖 They use real-world data to train neural networks for creating content that is more accurate than traditional effects or CGI.
• 🎭 The technology can be prompted by natural performances and superimpose faces, as demonstrated with a Tom Cruise impersonator.
• 🎤 The company showcased voice transfer technology, transporting a singer's voice onto another person's face in real-time.
• 🌐 The content creation is becoming easier and预示着 a future where individuals can be main characters in their own online content.
• 🔊 Live video manipulation is possible, moving from offline processing to real-time capabilities.
• 🎬 The technology raises ethical concerns about the misuse of deepfakes, such as creating fake news or non-consensual pornography.
• 🛡️ The speaker suggests raising public awareness about manipulated media and considering new legal rights for individuals over their data.
• 🌟 The technology has potential benefits, including emotional connections through hyper-realistic content and educational applications.
• 🏠 The technology could allow for virtual interactions with loved ones, preserving human experiences beyond their physical presence.

Q & A

What was the significant event that brought Metaphysic into the spotlight on the internet?

Metaphysic became prominent on the internet with the release of a deepfake video of Tom Cruise, known as 'Deep Tom Cruise,' which attracted around a billion views on TikTok and Instagram.

What does Metaphysic specialize in creating?

Metaphysic specializes in creating artificially generated content that looks and feels exactly like reality, using real-world data to train neural networks for highly realistic effects.

How does the technology work in terms of facial replication as demonstrated with Tom Cruise?

The technology uses AI to create a model of a person's face, which can then be superimposed onto another person's face in real time, creating a highly realistic and natural-looking replication.

Can Metaphysic's technology be applied to live video in real time?

Yes, Metaphysic's technology can be applied to live video, processing it so fast that it can be done in real time, as demonstrated during the TED conference.

What is the potential downside of this technology in terms of misinformation?

The technology can be misused to create deepfake videos that are indistinguishable from reality, leading to the spread of misinformation, fake news, and potentially damaging impersonations.

How does the speaker suggest we prepare for the challenges posed by deepfake technology?

The speaker suggests raising public awareness of manipulated media and considering new legal frameworks to empower individuals to own their real-world data and control the use of their photorealistic avatars.

What is the potential benefit of this technology in the entertainment industry?

The technology can be used to create highly realistic and engaging content, such as having a beloved actor like Tom Cruise appear in numerous films or creating new experiences that were not possible before.

How could this technology be used in education to enhance learning experiences?

An inspiring teacher could be virtually present in thousands of classrooms around the world, speaking every language, facilitating cultural exchange and socialization beyond traditional online learning platforms.

What emotional benefits could arise from interacting with photorealistic representations of loved ones?

Interacting with photorealistic representations can create a strong emotional connection, allowing users to relive experiences with loved ones or communicate with them in a way that feels very real and personal.

How did the speaker attempt to establish ownership over his own photorealistic avatar?

The speaker created a photorealistic avatar of himself and submitted it to the US copyright office in an attempt to establish legal ownership and control over its use.

What is the speaker's view on the future of this technology and its impact on society?

The speaker believes that the future will be a mix of weird and wonderful, with individuals needing to think differently about their identity and claim rights over their digital representations to navigate the new challenges and opportunities presented by this technology.

Outlines

00:00

🎬 AI-Generated Content: Tom Cruise Deepfake

The speaker discusses the prominence of their company due to the creation of a highly realistic fake Tom Cruise video that gained immense popularity on social media platforms. They express interest in creating a similar video for the Ted conference, highlighting the company's expertise in artificial content generation. The technology uses real-world data to train neural networks, creating hyper-realistic content. The co-founder's impersonation skills are also acknowledged, and the potential for voice transfer across different faces is demonstrated, showcasing the technology's ability to make anyone appear to speak any language naturally.

05:01

🕊️ Ethical Concerns and Creative Potential of Deepfakes

The conversation delves into the ethical implications and potential misuse of deepfake technology, such as creating false arrest photos of public figures or non-consensual pornography. The speaker acknowledges these concerns and emphasizes the importance of public awareness about manipulated media. Despite the risks, they argue that the technology is inevitable and will continue to advance rapidly. The discussion shifts to the creative possibilities, including the prospect of immortalizing actors like Tom Cruise in movies and creating emotional connections through realistic interactions with digital representations of loved ones. The potential for educational applications, such as having inspiring teachers in multiple classrooms worldwide, is also explored.

10:01

🛡️ Controlling Digital Selves and Owning Personal Data

The speaker addresses the issue of personal data ownership and the need for individuals to have control over their digital representations. They discuss the idea of decoupling human experiences from time and place, allowing for the creation of hyper-realistic media that can share the best aspects of our lives. The speaker shares their personal experience of creating a photorealistic avatar and attempting to copyright it, suggesting that owning such rights could allow individuals to control the misuse of their likeness on the internet. The conversation concludes with a reflection on the future being both weird and wonderful, with a need for new institutions and rights to navigate the challenges and opportunities presented by this technology.

Keywords

💡Deepfake

Deepfake refers to synthetic media in which a person's likeness is superimposed onto another body in a video, creating a convincing but fake appearance. In the video's theme, deepfakes are used to create realistic videos of celebrities like Tom Cruise, which raises concerns about authenticity and misuse. The script mentions a 'fake Tom Cruise video' that gained massive online attention, illustrating the power and prevalence of deepfake technology.

💡Artificially Generated Content (AGC)

Artificially Generated Content is media that is created using artificial intelligence, often to mimic real-world data. The video discusses AGC in the context of creating realistic and natural-looking content through neural networks, which is central to the company's work in producing deepfakes. The script highlights how AGC can be used to generate content that is indistinguishable from reality, leading to both creative and ethical challenges.

💡Neural Nets

Neural Nets, short for neural networks, are a set of algorithms designed to recognize patterns. They are foundational to the creation of AGC and deepfakes, as they learn from real-world data to generate realistic content. In the script, the company Metaphysic uses neural nets to create content that 'looks and feels exactly like reality,' showcasing the technology's advanced capabilities.

💡Real-time Processing

Real-time Processing refers to the ability of a system to process information immediately as it is received, without delay. The script mentions the advancement of deepfake technology to the point where it can be applied in real time, which is a significant technical achievement. This capability allows for live manipulation of video content, increasing the potential for both positive applications, like virtual performances, and concerning misuse.

💡Impersonation

Impersonation is the act of assuming the identity or characteristics of another person. In the context of the video, impersonation is achieved through deepfake technology, where one person's likeness and voice can be convincingly transferred onto another. The script humorously refers to the co-founder's ability to impersonate Tom Cruise, highlighting the realistic nature of the technology.

💡Voice Transfer

Voice Transfer is the process of applying one person's voice onto another person's video, which can be used to make it appear as if the person in the video is speaking in a different language or even singing. The script provides an example of this with Aloe Blacc's voice being transferred onto another person's face, demonstrating the versatility of deepfake technology in altering both visual and auditory aspects of media.

💡Authentic Media

Authentic Media refers to content that is genuine and not manipulated or altered. The video discusses the challenges of identifying authentic media in a world where deepfake technology can create highly convincing fakes. The script raises the issue of public awareness and the need for new institutions to verify authenticity, reflecting the broader implications of deepfake technology on trust and verification in media.

💡Cultural Exchange

Cultural Exchange is the sharing of cultural expressions or ideas between different cultures. The script suggests that deepfake technology could be used to facilitate cultural exchange by enabling a single teacher to teach in multiple languages simultaneously, thereby promoting understanding and interaction among diverse groups of students around the world.

💡Property Rights

Property Rights refer to the legal rights of an individual to own, control, and use their property. In the context of the video, property rights are discussed in relation to owning the data used to create deepfake representations. The script mentions the importance of individuals having control over their own data to prevent misuse, advocating for new legal frameworks to protect these rights.

💡Avatar

An Avatar, in the context of this video, refers to a digital representation of a person, often used in virtual environments or as a stand-in for the individual in digital media. The script discusses the creation of photorealistic avatars and the ethical considerations surrounding their use, including the potential for misuse and the need for individuals to have control over their own avatars.

💡Copyright

Copyright is a legal right that grants the creator of an original work exclusive rights to its use and distribution. The script describes an attempt to copyright an AI-generated version of oneself, which raises questions about the ownership and control of AI-generated content. This highlights the intersection of emerging technology and existing legal frameworks, and the need for updated regulations to address new forms of creative work.

Highlights

The company gained prominence with a deepfake Tom Cruise video that garnered a billion views on social media.

Metaphysic specializes in creating artificially generated content that appears real through neural network training.

AI technology can prompt natural performances and overlay faces, as demonstrated with a Tom Cruise impersonator.

The technology can also transfer singing voices, as shown with Aloe Blacc's voice onto another person's face.

Live video manipulation is possible, moving from offline to real-time processing.

A live demonstration was performed, superimposing Chris's face onto the presenter in real-time.

The technology raises concerns about the misuse of deepfake videos and the blurring line between reality and fiction.

The presenter discusses the importance of public awareness about manipulated media.

The technology's advancement is rapid and uncontrollable, with GPUs worldwide driving its development.

The potential benefits include entertainment possibilities, such as extending the presence of beloved actors like Tom Cruise.

Education could benefit from having inspiring teachers in classrooms worldwide, speaking multiple languages.

The emotional connection from hyper-realistic content can lead to more meaningful human interactions online.

The presenter suggests owning property rights over personal data to control the creation and use of photorealistic avatars.

A photorealistic avatar of the presenter was submitted for US copyright to establish ownership rights.

The future will require new ways of thinking about identity and claiming rights over personal digital representations.

The balance between the weird and wonderful aspects of this technology's future applications remains to be seen.

Transcripts

foreign

your company became prominent on the

internet with the release of a fake Tom

Cruise video deep Tom Cruise that I

think attracted like a billion views on

Tick Tock and Instagram which leads me

to my first question which is please can

we at Ted have our own Tom Cruise video

please uh you know I thought you might

ask so um a little earlier we we had a

crack maybe we'll have a look let's have

a look

what's up internet

I'm north of the Border a at the Ted

conference it's not short for Theodore

but nobody calls me Thomas so it's cool

it's Tom and Ted

yes uh Canada

seriously though everybody here very

nice very polite

especially the whales

I mean

that

so really at metaphysic we specialize in

creating artificially generated content

that looks and feels exactly like

reality so we take kind of real world

data we train these neural Nets and it

can more accurately than the effects or

CGI really create this content but looks

and feels so natural

um and so that is a great example of the

AI being kind of prompted by the natural

performance of a person and kind of the

face goes on top

and it helps the fact that your

co-founder is um you know he's a pretty

good Tom Cruise impersonation indeed

Miles Fisher is uh the foremost Tom

Cruise but not Tom Cruise

um I think you have another example as

well can we see that one yes absolutely

going kind of Beyond faces now talking

about voice

[Music]

[Music]

[Music]

tell us what's happening there so what

you see there is really it is the

singing voice of the lady singing

Spanish

um and aloe black who sings Wake Me Up

the Avicii song that he wrote with

Avicii um we are transporting her voice

across to his face so he doesn't sing in

Spanish and then suddenly we transport

it again

um her voice into his voice so anybody

in the future will be able to speak any

language it'll look Perfectly Natural

and this content is becoming more and

more easy to create and eventually it

will end up with a scale where we will

all be kind of main characters in our

own content on the internet

okay

[Laughter]

before we dig into that just a bit I

mean this what you've shown us

um recorded video there

um rumor has it you can also do this

with live video can that be right yes we

can do it live real time and this is

like really at The Cutting Edge of what

we can do today moving from offline

processing to we're processing it so

fast that you can do it in real time and

so well I'm going to challenge you and

your team to do a bit of a first yeah I

mean there's video of you right up on

that screen show us something surprising

you couldn't oh my gosh so wait so there

we go this is um you know a live

real-time model of Chris

um on top of me

um running in real time

and next you'll tell me that it can

oh Lord

I'm so uncomfortable with this

I am so uncomfortable Can it can it do

voice as well

um um we think they can we're we're

really pushing we're really pushing a

limit said AI technology now and I'm

talking exactly as I've heard

Scrambler is coming out is the one in

Italian for sensing I am I am I'm deeply

sorry everyone

subject to you to this you know that

there's something

possibly even worse so to come clean on

this they made me you know I did some

they took some shots of me a couple

weeks ago doing different facial

expressions that they captured a video

model and it turns out I discovered this

week that they can apply that not just

to Tom but to anyone

um and um so my my dear friend Sonny

Bates is here in the front row Sunny do

you consent to uh to channel inner me

for a minute well let's can we try that

I'm really worried about this

do we have sunny on screen

oh this they're Sunny oh no oh no oh no

[Laughter]

I

you know Chris you look amazing all

right enough of that cut that cut that

cut that right now I mean yes more Sunny

more of this Tom Tom Sunny Bates is the

woman who introduced me to Ted without

Sonny Bates none of us would actually be

here now and we do we reward her with

this I mean and now you've finally

become the master you know oh so look

okay amazing

um

it obviously occurs to everyone in this

room that there are some things that can

go horribly wrong with this

um and you know we've already seen you

know examples online of you know what

we've had photographs of trump being

arrested that were unreal there could be

video of it

um there's pornography that can use the

faces of celebrity all these things that

we've seen Deepak how how do you feel

about the downside of this technology so

personally you know we build this stuff

and I'm worried right worried is the

right Instinct for everybody to have and

then you know beyond that think about

you know what can we do to prepare

ourselves how can we try to impact the

future as it spirals in this direction

where as individuals it's going to be

kind of difficult to understand what's

computer generated and what's real and

so there are things that we can do there

raising public awareness of manipulated

media that's one that's you know this is

a great forum for that

will you claim to me that

if you were to shut down your company

right now it wouldn't stop the problem

of deep fake videos because the

Technologies out there that that's going

to happen anyway that that's not within

your control yes there are we're talking

about content that is so compelling if

you know we put any of ourselves inside

content and maybe it is talking to our

loved ones or just talking to our

friends on the beach and it's so

realistic that it looks real it's so

compelling everybody is trying to create

this content today all of the gpus in

the universe are driving it trying to

create this so it doesn't matter what

any one person does this will happen and

it's happening very very quickly

so I mean we'll talk about the upside in

a second but the but it seems like

we are going to have to get used to a

world

where we and our children will no longer

be able to trust the evidence of our

eyes

I think so

um we're going to have to understand a

new set of Institutions to verify what

is authentic media but then we can begin

to lead into some of the creative things

that happen from it and there are

benefits that come with that too so

it'll be an accommodation so talk a bit

about the uh the the benefits I mean

obviously on the entertainment side

there's an amazing number of

possibilities you have Tom Cruise we can

have impossible mission 273 in the year

21.50 right like he will be with us

forever we're working on number 75 000

one net right now yeah

I guess that I guess that is kind of

amazing like we love we've we love

lots of people in the world and we want

the possibility that with their

permission we can do more with them talk

about some other possible upsides what

we've seen in kind of building this

content and watching people interact

with it especially if they're kind of

interacting with themselves maybe it's

their younger 20 year old self or maybe

they're interacting with their partner

but the young version of their partner

there's this tremendous emotional

connection that comes from the very very

photorealistic Beyond The Uncanny kind

of content and so if we can start

deploying that

um among regular people if we can scale

it up so that it works for any kind of

person then we can begin to kind of you

know have more interesting meaningful

human kind of interactions and

relationships online and since the

pandemic we all spend more time online

but it's chat and it's email if we could

get more human emotion more feeling

there's a lot that we can do with that

right and so you know education is a

good example we could have an inspiring

teacher in thousands of classrooms

around the world speaking every language

in the world at the same time and

students could interact with each other

in a way that was you know goes beyond

Zoom there can be real cultural exchange

real socialization

there's a lot that we can do

Beyond you so that teacher example is is

powerful to me like like the fact that a

single teacher could turn a written

lesson into video in any number of

languages and extend indefinitely that

seems like a real amplification of

potentially of good you know human

intent

um I'm excited by that I still don't get

the family side of it like that like if

you want to have a human connection with

someone in your family isn't it like how

won't people just be creeped out by oh

my God I was just looking at your avatar

I thought I thought it was you you know

like how do we get isn't that just

creepy I I think this is definitely a

creepy element to this right and then

you go beyond that and the creepiness

drops away and the medium drops away and

it's the content and the connection

that's there so you know I imagine that

you know if I collect data from my

grandparents who are very old today then

in the future I'll be able to relive

experiences with them and communicate

with them and that it's not going to be

any good for them right they're probably

going to have passed on but for me it'll

help me process who I am my relationship

with them that idea of kind of

decoupling Human Experience both from

where it happens and the moment in time

that it happens I think that we can

create these experiences through hyper

real photorealistic media that allow us

to share the best of our experiences the

best of who we are I'm going to be very

curious to see who's who can feel that

right now my guess is that there's going

to have to be lots of experiments and a

lot of things are going to creep us out

and maybe we'll find some things that

are just absolutely incredible but I

have an important question for you which

is how the hell do I control me now hour

like you've got me on video what's to

stop my me being misused across the

internet now that's right I think that

as we kind of allow companies to create

these realistic experiences as

individuals we need to be empowered to

own our real world data the data used to

train the algorithms and we need to

control how our photorealistic avatars

are created and where they're used so to

this extent I was looking at kind of

conventional current legal institutions

to see what we could do to create new

rights so I created a photorealistic

avatar of myself submitted it to the US

copyright office to see if they would

register my copyright in it and this is

what the video looked like

here is the AI realistic version of

myself

even if the appearance of this AI

representation of myself may change

cosmetically or if I change my hair or

add creative features my intention is to

create this AI version of myself that

embodies the essence of who I am as a

person under any circumstance

wow

so I think you have shown us what is

going to be a repeated theme in this

conference which is the future is going

to be

weird and wonderful at the same time

quite what the balance is between those

two TBD but this is a world where each

of us is going to have to think

differently about who we are and claim

these rights what you're saying is that

if people have this right you can

picture a world where for example if a

video goes viral on YouTube without your

consent you'll be able to take it down

because of some link back to this that's

right I think the most important thing

that we can do when we're talking about

data from a real world being able to

power these things is that we need to

own property rights over the data we

shouldn't sign it over to companies

through terms of service we shouldn't

give it away if you fundamentally own it

then you can be in control and you're at

the ground level of all of the economies

and all of the use cases that are going

to spin up through history

it's a lot but uh we're on the way now

Tom Graham thank you so much for sharing

this technology thank you very much

and sunny Bates I'm so sorry thank you

so much