How a Deepfake Almost Ruined My Political Career | Cara Hunter | TED

This talk contains mature language

“You’re a little whore,

and we've all seen your little video."

That was the text message that was sent to me in April of 2022.

I'm sitting in my grandmother's living room

in what is her 90th birthday,

surrounded by family and friends

as my phone blows up with messages from strangers

right across the country

who say they have seen a video of me

engaging in hardcore pornographic activity with a man.

I knew this was impossible.

With just three weeks out from my election,

I felt as though my career was crumbling before my very eyes.

My heart pounded,

my mind raced, sweat beaded on my skin.

And then I watched the video, and my worst fear was realized.

Although this woman in the video was not me,

she looked exactly like me.

Impossibly like me.

Eerily like me.

I had so many questions running through my mind.

Was this AI?

Was it not?

Who made this?

How did they make it?

Why did they make it?

So I did what anyone would do,

and I approached my local police service to ask for advice, for guidance.

And really, where did I go from there?

But they informed me that they wouldn't have the cybercrime technology to assist,

to find out where this video came from.

And it was from that moment I knew that I was on my own.

Now to set the stage, as you can probably tell,

I’m from Ireland,

and to be exact, I'm from Northern Ireland,

which is an even smaller place.

We have just 1.8 million people,

very similar to the size of Vienna.

So you can imagine a rumor of this sinister nature,

particularly in the world of politics, can go very far, very fast.

And that old saying, “seeing is believing” began to haunt me.

And in the weeks leading up to my election,

this video, this false video,

was shared thousands and thousands of times across WhatsApp.

And attached to this video was photos of me at work,

smiling, campaigning,

building a sense of trust with my constituents.

And as the weeks went on,

messages flooded in faster and faster,

and they were of a very vile and sexual nature.

Ding!

"We've all seen your little video."

Ding!

"You should be ashamed of yourself."

Ding!

"Ah, now I see how you got your position in politics."

It was very difficult.

And having been in politics since the age of 23,

and at this point I've been in it for about four to five years,

and I'm from Northern Ireland, which is a post-conflict society,

still very deeply divided.

So I anticipated challenges,

I anticipated disagreements,

I even anticipated attacks, is politics after all.

But what I did not anticipate was this moment.

This was different.

This was the moment where misogyny meets the misuse of technology,

and even had the potential to impact the outcome of a democratic election.

And the sad thing for me was this lie became so far spread,

so far, so fast,

that even my own family started to believe it.

Some would say that they'd heard it at a golf club,

others would say they heard it at the bar,

and of course, some even said they heard it in a locker room.

A really good example of how far this went

was people that I knew my entire life

would pass me in the street without whispering a word.

People like school teachers,

people I had a sense of trust with

and, you know, an affinity with.

And that was really hard.

It felt like overnight I was wearing a scarlet letter.

And as things moved on

and were about two, three weeks out from the election,

I kept receiving messages, and it got wider and wider.

It was global.

Not only was I receiving messages from Dublin and from London,

but I was also receiving messages from Massachusetts, Manhattan,

and I was getting so many follows on my political social media,

predominantly from men hungry for more of this scandal.

And this intersection of online harms impacting my real life

was something I found utterly strange and surreal.

But it got to the point where I was recognized on the street

and approached by a stranger who asked me for a sexual favor.

And it was just, for me, it was like in the blink of an eye,

everything had just changed,

and it was utterly humiliating.

I didn't want to leave the house,

and I had turned notifications off in my phone

just so I could kind of catch my breath,

but this wasn't ideal in the lead up, of course, to an election.

And for me, I think that was the purpose of this false video, was to do just that.

But what hurt the most for me

was sitting down my father

and having to explain to him this strange, surreal situation.

My father is an Irishman,

completely disconnected from tech,

and so having to explain

this horrific situation was an entire fabrication

was very hard to do.

This was this strange moment

where the online world met my life, my reality.

Not only having the impact to ruin my reputation,

but have the capacity to change the outcome

of a democratic election.

And, you know,

for years I spent so much time building trust with my constituents.

I mean, we all know how much people like politicians.

You know, we’re as likeable as the tax man.

So for me, it was hard.

It was really hard

because it was years of hard work.

You know, I’m so passionate about my job,

and this video, this complete falsehood,

had the ability to just undermine years of hard work in mere seconds.

But instead of succumbing entirely to victimhood,

I ask myself today,

you know, where do we go from here?

And how can AI evolve to prevent something like this happening again?

Not only has it happened to me,

but we want to future-proof and ensure that this doesn't happen

to the women of tomorrow.

How can we, you and I, people who care about people,

ensure that this is a tech for good?

How can we, the policymakers, the creators, the consumers,

ensure we regulate AI and things like social media,

putting humans and humanity at the center of artificial intelligence?

AI can change the world.

In fact, as we've heard today, it already has.

In a matter of seconds,

people who speak completely different languages

can connect and understand one another.

And we've even seen the Pope as a style icon in a puffer jacket.

So some really important uses right there.

But then in my case as well,

we can also see how it is weaponized against the truth.

And good examples of this would be art that appears like reality,

AI-generated reviews unfairly boosting certain products

and things like chatbot propaganda.

And then, politically speaking,

we've seen over the years deepfakes of Nancy Pelosi slurring,

Joe Biden cursing

and even President Zelensky

asking his soldiers to surrender their weapons.

So when AI is used like this, to manipulate,

it can be a threat to our democracy.

And the tech is becoming so advanced

that it's hard to differentiate fact from fiction.

So how does AI interfere with politics?

And for us as politicians, what should we be worried about?

Could truth and democracy become shattered by AI?

Has it already?

Well, to dive a little deeper here,

I think firstly we need to talk about the concept of truth.

Without truth, democracy collapses.

Truth allows us to make informed decisions,

it enables us to hold leaders accountable, which is very important.

And it also allows us, as political representatives,

to create a sense of trust with our citizens and our constituents.

But without that truth,

democracy is vulnerable to misinformation,

manipulation, and, of course, corruption.

When AI erodes truth,

it erodes trust,

and it undermines our democracy.

And for me, in my experience with a deepfake,

I've seen what a fantastic distortion tool that deepfakes really are.

So how can we safeguard democracy from this ever-advancing technology?

It's becoming ever harder to distinguish between real and synthetic content.

And politically, what role does AI play in the future?

And I can't talk about politics without talking about media as well.

They're undeniably linked, they're intertwined.

And I think journalism has its own battle here as well.

From AI algorithms boosting articles unfairly

to clickbait headlines,

and then also moments where they can manipulate the public as well.

But politically speaking,

we've seen AI-tailored political messaging

influencing voters,

we’ve seen it adding to existing bias.

And definitely I think we all have that aunt that’s on Facebook

and kind of believes anything.

So for me, as a politician,

I think it's really important we dive a little deeper

into the relationship of AI, journalism and media.

But it also puts us at risk

of creating a more divided and reactionary society,

because falsehoods can create a lot of reaction.

And for myself, coming from Northern Ireland,

which is that post-conflict society,

I do have concerns about how it could shape our political landscape

and other places across the globe.

Sadly, this deepfake video

is not the only instance of me having experienced abuse with AI.

Just six months ago,

I received 15 deepfake images of myself in lingerie

posing provocatively.

And I thought to myself, here we go again.

And you know, I spoke with some other female politicians.

Thankfully, where I represent, we have more women getting into politics.

But I had a really good conversation with them,

and it’s around, if this is happening to you now,

what happens to me tomorrow?

And I think this really strikes at the heart of the climate of fear

that AI can create for those in public life.

And I don't blame women.

It's very sad.

I don't blame women for not wanting to get into politics

when they see this kind of technology come forward,

so that's so important that we safeguard it.

What also concerned me was the position of elderly voters

and perhaps their media literacy,

their comprehension of this technology.

People who don't use the internet,

who perhaps are not aware of AI and its many, many uses.

So that was really concerning as well.

But it doesn't have to be this way.

We can be part of the change.

For me and my video,

I still don't know, to this day, who did this.

I can imagine why they did it, but I don't know who did it.

And sadly for me and for across the globe,

it still wouldn't be considered a crime.

So from being the enemy of fair, transparent,

good-natured politics,

to warfare, to international interference,

despite its beauty and its potential,

AI is still a cause for concern.

But it doesn't have to be this way.

I feel passionately that AI can be a humanistic technology

with human values

that complements the lives that we live

to make us the very best versions of ourselves.

But to do that,

I think we need to embed ethics into this technology,

to eliminate bias, to install empathy

and make sure it is aligned with human values and human principles.

Who knows, our democracy could depend on it.

And today I know we have some of the brightest minds in this room.

I heard some of the talks earlier, and I'm certain of that.

And I know each and every one of you have weight on your shoulders

when looking to the future.

But I also know each and every one of you want to see this tech for good.

And what gives me hope

is witnessing the movement right across the globe

to see this hard journey begin,

to regulate this ever-advancing technology,

which can be used for bad or for good.

But perhaps we, each and every one of us in this room,

can be part of finding the solution.

(Applause)