Russia and Iran use AI to target US election | BBC News

Summary

The transcript from 'AI Decoded' discusses the threat of generative AI in spreading disinformation, with a focus on deepfakes and their impact on democracy and elections. It covers California's new law against deepfakes in elections, watermarking as a potential solution, and the role of social media companies in regulation. Experts weigh in on the challenges of detecting deepfakes and the importance of critical media literacy. The show also features AI's role in debunking conspiracy theories through chatbots, highlighting the potential of AI in combating the spread of false information.

Takeaways

• 📜 California has passed a bill making it illegal to create and publish deep fakes related to upcoming elections, with social media giants required to identify and remove deceptive material from next year.
• 🐱 The script discusses the problem of AI-generated memes, like those of cats and ducks, which have fueled rumors with dangerous consequences.
• 🏛️ Beijing is pushing for AI to be watermarked to help retain social order, placing responsibility on creators to ensure the authenticity of AI-generated content.
• 🎤 The script mentions how AI has been used to hijack the image of celebrities like Taylor Swift, who was falsely shown endorsing a political candidate.
• 🌐 The Microsoft Threat Analysis Center in New York City works to detect and disrupt cyber-enabled influence threats to democracies worldwide.
• 🔍 The center has detected attempts by Russia, Iran, and China to influence the US election, with each nation using different tactics such as fake videos and websites.
• 🤖 AI is being used to combat the spread of misinformation, with researchers developing tools to detect deep fakes and provide explanations for their authenticity.
• 💡 The discussion highlights the need for watermarking as a potential solution to identify genuine content, but also acknowledges the challenges in keeping up with advancing technologies.
• 🌐 There's a call for a global approach to traceability in AI-generated content, similar to supply chain management, to ensure the origin and authenticity of digital creations.
• 🤖 The script introduces a chatbot designed to deprogram individuals who believe in conspiracy theories by engaging them in fact-based conversations.

Q & A

What is the significance of the bill signed by Governor Gavin Newsom in California regarding deep fakes?

The bill signed by Governor Gavin Newsom makes it illegal to create and publish deep fakes related to upcoming elections. Starting next year, social media giants will be required to identify and remove any deceptive material, marking California as the first state in the nation to pass such legislation.

How does generative AI amplify the threat of disinformation?

Generative AI tools, which are largely unregulated and freely available, have the potential to create convincing fake content, including deep fakes and manipulated media, which can be used to spread disinformation and undermine trust in elections and freedoms.

What is the role of the Microsoft Threat Analysis Center in New York City?

The Microsoft Threat Analysis Center, located in New York City, is a secure facility that monitors attempts by foreign governments to destabilize democracy. It detects, assesses, and disrupts cyber-enabled influence threats to democracies worldwide.

How do the analysts at the Microsoft Threat Analysis Center detect foreign influence attempts on US elections?

Analysts at the Microsoft Threat Analysis Center detect foreign influence attempts by analyzing data and reports, identifying patterns, and advising governments and private companies on digital threats. They have detected simultaneous attempts by Russia, Iran, and China to influence the US election.

What challenges do AI tools face in detecting deep fakes?

AI tools face challenges in detecting deep fakes due to the continuous advancement of generative AI technologies, which can create increasingly realistic fake content. Additionally, AI tools may struggle with images or videos that are too far away from what they have seen during training, leading to potential misclassifications.

What is the potential solution to the deep fake problem discussed in the script?

One potential solution discussed is the use of AI to detect misinformation and deep fakes. This involves training AI tools to identify inconsistencies and anomalies in content, and providing explanations for why certain content is flagged as a deep fake.

Why is watermarking proposed as a solution to the deep fake problem?

Watermarking is proposed as a solution because it can provide a form of traceability and authenticity to digital content. It would allow for the identification of original and verified content, helping to distinguish it from deep fakes.

How does the concept of 'situational awareness' relate to the detection of deep fakes?

Situational awareness in the context of deep fake detection refers to the ability to proactively monitor and analyze content on social media platforms using AI tools. This allows for the establishment of a global scale understanding of where and when disinformation is being spread.

What is the 'debunk bot' mentioned in the script and how does it work?

The 'debunk bot' is an AI chatbot designed to converse with conspiracy theorists using fact-based arguments to debunk their beliefs. It draws on a vast array of information to engage in conversations and has shown success in reducing conspiracy beliefs by an average of 20% in experimental settings.

How does the debunk bot approach the challenge of changing deeply held beliefs?

The debunk bot approaches the challenge by providing tailored information and facts directly related to the specific conspiracy theories that individuals hold. It engages in a conversation that summarizes and challenges the beliefs, using evidence to persuade users away from their conspiracy theories.

Outlines

00:00

📜 AI and the Threat of Disinformation

The segment begins with a discussion on the challenges posed by generative AI in creating and disseminating disinformation. It highlights the recent legislation in California that criminalizes the creation of deep fakes related to elections and the upcoming requirement for social media platforms to identify and remove deceptive content. The conversation also explores the global impact, including China's efforts to watermark AI content and the manipulation of public figures like Taylor Swift. The segment emphasizes the role of AI in both creating and combating disinformation, with a focus on the importance of critical media literacy and the challenges of regulating technology across borders.

05:00

🔍 Deep Fakes: Detection and Regulation Challenges

This paragraph delves into the difficulty of detecting deep fake imagery and audio, and the potential solutions such as watermarking to authenticate content. It discusses the ongoing 'whack-a-mole' game with technology, where advancements in detection are met with new methods of manipulation. The conversation includes the role of social media companies in regulating content, the challenges of enforcing regulations globally, and the importance of preparing citizens with critical media skills. It also touches on the debate over First Amendment rights and free speech in the context of regulating AI-generated content.

10:02

🕵️‍♂️ AI Tools for Detecting Deep Fakes

The focus of this section is on the development of AI tools to detect deep fakes. It features a discussion with Dr. Christian Schroedl from the University of Oxford, who is researching methods to identify deep fakes using AI. The conversation includes the use of AI to track down AI-generated misinformation, the limitations of current detection tools, and the need for further research. Examples of deep fake images, such as the Pope in a puffer jacket, are used to illustrate the challenges in detection and the potential of AI in providing explanations for why certain images are deemed deep fakes.

15:04

🌐 Social Media and the Spread of Disinformation

This segment discusses the role of social media in the spread of disinformation and the vested interest companies have in maintaining a reliable information ecosystem. It highlights the potential of AI to not only create but also combat deep fakes and misinformation. The conversation includes the potential of context-aware AI to determine the authenticity of content and the importance of situational awareness on a global scale. The segment also touches on the challenges of disinformation spread by conspiracy theories and the difficulty of changing deeply held beliefs with facts alone.

20:07

🤖 Debunking Conspiracy Theories with AI

The final paragraph introduces the concept of using AI chatbots to debunk conspiracy theories. It features an interview with Dr. Thomas Costello, who discusses the development of a 'debunk bot' that engages with conspiracy theorists using fact-based arguments. The segment covers the effectiveness of the chatbot in reducing belief in conspiracy theories, the potential for incorporating such technology into existing platforms, and the challenges of changing beliefs that are deeply ingrained. It also raises the risk of using AI for disinformation if not properly regulated.

25:08

🏁 The Role of Industry in AI Regulation

In the concluding remarks, the discussion turns to the role of the industry in self-regulating AI technologies. It highlights the importance of continuous improvement in AI models to prevent the spread of false content and the business imperative for companies to provide accurate and reliable information. The segment emphasizes the speed at which industry can adapt compared to legislation and the potential for industry-driven solutions to lead the way in addressing AI-related challenges.

Keywords

💡Deep fakes

Deep fakes refer to synthetic media in which a person's likeness has been manipulated or replaced using AI. In the video, deep fakes are discussed as a significant threat to disinformation, particularly in the context of elections. The script mentions California's legislation against deep fakes related to elections, highlighting the urgency and real-world implications of this technology.

💡Disinformation

Disinformation is the deliberate spread of false information to deceive and mislead. The video script explores how generative AI tools can amplify disinformation, undermining elections and freedoms. It is a central theme, with examples including AI-generated memes that fueled rumors with dangerous consequences.

💡Generative AI

Generative AI refers to AI systems that can create new content, such as text, images, or videos. The script discusses generative AI as tools that, while creative, can also be misused to generate disinformation, emphasizing the dual-use nature of this technology.

💡Watermarking

Watermarking in the context of AI refers to the practice of embedding a digital signature or mark into content to verify its authenticity. The script suggests watermarking as a potential solution to identify and combat deep fakes, though it also acknowledges the challenges in implementing such measures.

💡Election interference

Election interference is the manipulation or disruption of an electoral process. The video discusses how foreign governments use AI to create deep fakes and other forms of disinformation to interfere with elections, as detected by the Microsoft Threat Analysis Center.

💡Cyber-enabled influence threats

Cyber-enabled influence threats are attempts to sway public opinion or disrupt social order through digital means. The script mentions the Microsoft Threat Analysis Center's role in detecting and disrupting such threats, which include foreign attempts to influence US elections.

💡AI detection

AI detection refers to the use of AI algorithms to identify deep fakes and other forms of manipulated media. The script features a discussion on how AI can be used to track down deep fakes, suggesting that AI can be a tool for combating the very problems it helps create.

💡Digital threats

Digital threats encompass a range of online risks, including cyberattacks, data breaches, and disinformation campaigns. The video script discusses how organizations like the Microsoft Threat Analysis Center advise governments and companies on these threats, highlighting the importance of digital security in the modern era.

💡Provenance

Provenance, in the context of digital content, refers to the ability to verify the origin and history of a piece of content. The script discusses the importance of establishing provenance for AI-generated content, suggesting that cryptographic signatures or other forms of traceability could help authenticate content.

💡Conspiracy theories

Conspiracy theories are explanations of events or situations that invoke secret plots by powerful individuals or groups. The video script mentions the use of AI chatbots to debunk conspiracy theories by providing tailored facts and evidence, aiming to reduce belief in such theories.

💡Fact-based arguments

Fact-based arguments are persuasive statements grounded in verifiable facts. The script describes an AI chatbot designed to engage with conspiracy theorists using fact-based arguments to challenge and potentially change their beliefs, showcasing the potential of AI in promoting accurate information.

Highlights

California Governor Gavin Newsom signed a bill making it illegal to create and publish deep fakes related to upcoming elections.

Social media giants will be required to identify and remove deceptive material from next year in California.

The challenge of distinguishing between fake and real content is becoming increasingly difficult with generative AI tools.

AI memes of cats and ducks have fueled rumors with dangerous consequences, illustrating the impact of disinformation.

Beijing is pushing for AI to be watermarked to retain social order in a world of manipulated messages.

Taylor Swift was hijacked by the former president who shared fake images of her fans endorsing him, showing the personal impact of AI disinformation.

Microsoft's Threat Analysis Center in New York is at the forefront of defending against cyber-enabled influence threats to democracies.

Russian, Iranian, and Chinese attempts to influence the US election have been detected simultaneously for the first time.

The US election's dramatic nature is complicating outside interference, particularly affecting Russian strategies.

Iranian election influence activity has been detected via bogus websites, currently under FBI investigation.

China is using fake social media accounts to provoke reactions in the US public, increasing hostility on social media.

The debate over watermarking as a solution to generative AI disinformation is discussed, with concerns about potential manipulation.

The importance of preparing citizens with critical media skills to construct narratives and verify information sources is emphasized.

The challenge of regulating AI and deep fakes globally, especially when companies may relocate to less regulated countries, is highlighted.

The potential for AI to track down deep fakes and misinformation is explored, with AI being used to solve the problems it creates.

Researchers are developing AI tools to identify deep fakes by creating explanations for why content is identified as fake.

The limitations of current AI tools in detecting deep fakes, such as issues with finger detection and temporal inconsistencies, are discussed.

The potential for AI to rearchitect systems to provide a cryptographic signature of authenticity for content is considered.

The idea of using AI chatbots to deprogram conspiracy theorists by providing fact-based arguments is introduced.

The debunk bot, an AI chatbot, has shown success in reducing belief in conspiracy theories by 20% on average in experimental conversations.

Transcripts

you are watching the context with me

Christian Fraser it is time for our

regular Thursday feature AI

[Music]

decoded welcome to the program freely

available largely unregulated the

creative tools of generative AI now

amplifying the threat of disinformation

how do we tackle it what can we trust

and how are our enemies using it to

undermine our elections and our freedoms

this week Governor Gavin Nome signed a

bill in California that makes it illegal

to create and published deep fakes

related to the upcoming election and

from next year the social media Giants

will be required to identify and remove

any deceptive material it is the first

state in the nation to pass such

legislation is it the new Benchmark some

of this stuff obviously fake some of it

deigned to poke fun but look how these

AI memes of cats and Ducks powered the

pet eating Rumor Mill in America with

dangerous

consequences it is a problem too in

China how does the Communist Party

retain social order in a world where the

message can be manipulated Beijing is

pushing for all the AI to be watermarked

and he's putting the onus on the

creators and from politics to branding

there is no briger brand than Taylor

Swift hijacked by the former president

who shared fake images of her fans

endorsing him it affects us

all with me as ever in the studio our

regular commentators and AI presenters

Stephanie ha is here and from Washington

our good friend Miles Taylor who worked

in National Security advising the former

Trump Administration we'll talk to them

both in a second but before we do that

we're going to show you a short film one

of the many false claims that has

appeared online in recent months was a

story that Cara Harris had been involved

in a hit and run accident in 2011 that

story was created by a Russian troll

farm and was one of the many

inflammatory stories Microsoft

intercepted the threat analysis unit

that does their work in New York is at

the very Forefront in defending all our

elections our AI correspondent March

schlack has been to see it Time Square

New York City an unlikely location for a

secure facility which monitors attempts

by Foreign governments to destabilize

democracy it is however home to mtag the

Microsoft threat analysis Center its job

is to detect assess and disrupt cyber

enabled influence threats to democracies

worldwide the work that's carried out

here is extremely sensitive we the very

first people that have been permitted to

film

inside it's also the first time Russian

Iranian and Chinese attempts to

influence the US election have all been

detected at once all three are in play

and this is the first cycle where we've

had all three that we can definitely

point to individuals from this

organization serve on a special

presidential Committee in the Kremlin

advis reports compiled by these analysts

advise governments like the UK and us as

well as private companies on digital

threats this team has noticed that the

dramatic nature of the US election is

complicating attempts at outside

interference the biggest impact of the

switch of president uh Biden for vice

president Harris has been it's really

thrown the Russians so far off their

game they really focused on Biden as

somebody they needed to remove from

office to get what they wanted in

Ukraine Russian efforts have now pivoted

to undermining the Harris Waltz campaign

via a series of fake videos designed to

provoke

controversy these analysts were

instrumental in detecting Iranian

election influence activity via a series

of bogus websites the FBI is now

investigating this as well as Iranian

hacking of the Trump campaign we found

that in the source code for these websit

they were doing was using AI to rewrite

content from a real place and using that

for the bulk of their website and then

occasionally they would write real

articles um when it was a very specific

political point they were trying to make

the third major player in this election

interference is China using fake social

media accounts to provoke a reaction in

the US public experts are unconvinced

these campaigns affect which way people

actually vote but they worry they are

successful in increasing hostility on

social media Mark chisl BBC News yeah

that gives you an idea of just how quick

this is advancing Stephanie do you do

you think we're almost at a point as the

technology improves the creative

technology that we're going to be very

close very soon to not knowing the

difference between fact and fiction it's

getting harder and harder to detect a

lot of the deep fake imagery audio is

particularly very difficult to detect

it's a lot easier to fake so yes I think

we're right now possibly in the last us

election where it's kind of easy to see

when you're being manipulated and the

the trick really is do you want to

believe it because what this is all

about is really hijacking your emotions

and watermarks because that is often the

the goto solution to this why would that

not be the the answer to all the ills of

generated generative AI I still wonder

if there would be ways of manipulating

even that but it's probably a pretty

good start it's just that thing you

always feel like you're playing

whack-a-mole with these Technologies you

know you do one thing and then it

advances and you have to catch up again

so we would probably start with

watermarks and then there would be an

advance and a kickback and we'd have to

react to that and so on and so forth I

think it's also about preparing citizens

though to have the critical media skills

that we all need to be able to construct

narratives look at who is giving us

information and just does it check with

reality miles um I was saying to

Stephanie this is a good step forward

what's happened in California this week

you've got the governor there putting

the onus on the social media companies

and on the creative companies to do

something about this and particularly

around the election and then Stephanie

said to me well okay American companies

regulated by American legislators why

wouldn't they just go to

China look I mean I think that's one of

the concerns always when it comes to

Tech regulation and and Christian you

remember the debate well over encryption

in the United States there was the San

Bernardino terrorist attack uh you know

almost 10 years ago now where the FBI

could not get into the shooter's phone

and it led to a big debate in the United

States about these encrypted messaging

apps like Telegram and signal and

whether it should be legislated that

those were forbidden in the United

States opponents of those laws though

said well sure you can outlaw them here

but someone overseas is going to create

the same apps and it's going to be

really difficult to prevent people from

using a version of it overseas we Face

the same problem here with regulations

around deep fake deep fakes and AI it's

only as far as us legislation and law

enforcement can reach that those types

of things can be enforced so there is a

big challenge here but also there's a

domestic challenge about the first First

Amendment implications and Free Speech

implications and of course Governor

nome's signing of that law has opened up

that debate as well so there will be a

lot of contention the next few years

about how to get this right from a

legislative and Regulatory standpoint

the other thing that occurs to me and we

talk about protecting Children online

all the time on this program one of the

issues the companies always come up

against is finding the material and

getting rid of it if you are having to

find very good deep fake material

that process becomes much more difficult

doesn't it and how do we find a metric

to to hold the social media companies

and the online companies to

task well I think Stephanie said

something really important here which

was the game of whack-a-mole you're

playing if you think that watermarking

you know basically sticking a putting a

sticker on this content and saying this

is fake if you think that's a solution

it's going to be really hard to keep up

a lot of the experts I talk to in AI say

that maybe that a short-term solution

but in the longer term you have to

rearchitecturing

at this place at this time and that

can't be changed right it's tied to a

public Ledger uh not that people can see

your photos publicly but that's a

cryptographic signature that can't be

broken eventually all of our Tech will

be signed with that Providence that says

I am real and you'll know if it's not

real because it won't have that point of

creation certification but it's years

before we're there and in the meantime a

lot of difficult conversations are going

to be had it's almost a supply chain

approach or even a criminal approach

when you have a chain of evidence and

you have to be able to follow it all the

way through and you can't tamper with it

or when we had mad C disease here in the

United Kingdom many years ago people

suddenly wanted to know when they were

going grocery shopping they wanted to

buy some beef what farm did it come from

and suddenly people realized they needed

traceability all the way through the

food chain so I'm wondering if there's a

parallel there to help people understand

all of the things that you're creating

can have that encoded so you would

always be able to know it's like

following through like a painting when

is a painting sold it might go through

50 different hands if it's 400 years old

you know before it finally ends up in

the Met um where did it come from was it

illegally bought you know Etc you you

should be able to follow data through in

the same way let's bring in someone uh

who is working in in this field here in

at the studio with us is Dr Christian

SCH deit uh he is a senior research

associate in machine learning at the

University of Oxford he and his team are

researching how to identify some of

these deep fakes using AI welcome to the

program um we were just talking about

how quickly things are advancing to the

point where to the naked eye it's

becoming more difficult certainly with

imagery what sort of Technology are you

developing that makes that easier yes so

Christian um I really like this

discussion um I think um the solution to

our problems of establishing Provence um

of content um will involve both a lot of

research but also wider adoption of

existing Technologies so in terms of

research I think the clip really brought

home you know that um AI is being used

to amplify the misinformation problem so

let's use AI to solve it so some of the

research that I do is about using AI to

detect misinformation so you're using

the AI to track down the Deep fake AI so

basically yes so so what I did the

summer spending um you know doing some

research doing some research with BBC

verify um and University of Oxford was

um just you know when you have a picture

for example um explain

whether it is a deep fake or not let's

bring one up I've got one um that I

think you've looked at and people will

be familiar with this it's it's the Pope

in a puffer jacket which actually did

get into some uh news streams around the

time that this photo came out so

although we're joking it did actually

deceive quite a lot of people show me

what you did with this yeah so exactly

so you can see um the pope and the

puffer jacket obviously from the context

it's quite clear it's a deep fake right

um and it's probably for entertainment

purposes but a human expert for example

BBC verify could look at this picture

and could um find the details that are a

bit off for example the spectacle seem

to be fused into the cheeks or the

crucifix doesn't quite attach to the

chain right and so the question is um

you see it's very important to have

these explanations as well not just like

a number of like this is 0.7% or 0.7

deep fake or not but you need to have an

explanation for why it is a deep fake so

we now have ai tools that can create

these explanations as well right um what

something that you put on the desktop

something that you could run a

photograph through yeah potentially yes

okay um but these tools still have a lot

of failure cases and this is where we

need more research okay um yeah where do

they fail and why famously it's things

like they can't get fingers right so you

might get six fingers on a hand yes so

this is a classic on videos for example

right like you have some sort of

temporal inconsistency so an object

disappear suddenly for example right um

but the problem is that these tools are

trained on a lot of data and they're

learning so-called features patterns

right that help them to make these

decisions now it can happen that

sometimes these patterns are present in

some images that are too far away from

what it has seen during getting too

technical on that can you explain that

to people is it is it a pixel difference

is it is it in the way I mean it's not

in the way the image looks is it they're

looking the the AI presumably is looking

deeper into the image than that yes so

the AI is actually taking in an image

and then it is um projecting this into

some very high dimensional space um and

within this High dimensional space um

basically um um you then do like a

dimensionality reduction into a lower

space and then in this lower space um

what you can do is um you can um uh form

these features okay and then if you have

an image that it hasn't seen during

training then these features um might

not generalize to that image and then

like you can have issues where um an um

evokes some Impressions that that are

that are wrong right so you see some

Reflections or something and actually

they are not deep fake but the Stephanie

Stephanie mentions the photographs that

they they struggle with I've got one

here this is lonel Messi kissing the

World Cup much to my shine but um but

the but the um this one is

real but the machine thought it was fake

why yes so um so the machine might think

so just because it maybe hasn't seen an

image that's close enough to this

picture in its training set right and as

we always get new images in for example

winning the World Cup Messi winning the

World Cup was a new occasion um um it

might think that for example some

Reflections and the trophy um or the way

Messi um holds his hand um or maybe the

skin tone um aren't natural and the

problem is we then get these

explanations and these explanations can

be very very convincing um but they're

nevertheless wrong miles do you like

this idea of AI tracking

AI deep

fakes I don't just like it Christian I

love it we've got to use AI against AI

to protect ourselves it's actually going

to be our best asset and one of the

things that's interesting that's

happening right now is we always focus

on who's developing the technology that

could be used for bad but uh my fellow

oxonian there on Set uh and and a lot of

folks around the world are now investing

time and resources into building

companies on deep fake detection I mean

there are companies in the United States

like true pick and reality Defender that

are exciting they're venture-backed a

lot of people want to go work for them

and what do those companies do they

focus solely on trying to prove what is

and isn't real and one of the things

that's just become possible really only

in the past few months is some of these

Technologies are leveraging context

awareness of the world to determine

whether something's fake or real so

these models aren't just looking at the

image and saying it looks manipulated

the models can also say well the Pope

the past couple weeks has been on

vacation in Italy there's no way this

photo was just taken and he was wearing

a puffer jacket and they can give you a

confidence score and that's exciting are

you incorporating that in your in your

abely so this is incorporating wider

context on where the content is found

and when it is found and who is depicted

so sort of semantic information

absolutely yes it strikes me that the

social media companies and the online

companies have a vested interest in this

because if you can't tell fact from

fiction you get what's called a liar's

dividend right that that actually you

become a disruptor you you you poison

the we so much that actually No One

Believes anything and that's not good

for a social media model that makes

their money from from spreading news and

and informing people when it just raises

the question of what social media is for

right so it was quite exciting at first

when it was this new thing and you could

stay in touch with your friends and then

a lot of people journalists would use

certain tools to keep up with the news

and get breaking news fast but once that

starts feeling like actually they're

just reading your data or you're looking

for news to get it fast but it's not

actually reliable and it's being flooded

information ecosystems being flooded all

the time eventually you might just turn

off and that's without even going into

the mental health implications of being

on these sites right which we know are

really harmful for people so I wonder

sometimes if we might be having lived

through the Golden Age of social media

and we're now entering this new phase

and if it isn't cleaned up people could

just end up leaving it or only going to

the way that you would read the national

Inquirer in the United States to read

about aliens or something are the big

developers interested in what you're

doing yes absolutely so so the summer my

collaboration was with a big tech

company in fact um so there is a lot of

interest in these Solutions actually the

interest goes even further so what we

can do now we can proactively try to

look for deep fakes and disinformation

in social media platforms right using

autonomous agents so I think this is

where things are going and then we can

establish this situational Awareness on

a sort of global scale H which miles

also I've got to also ask you is this

the right environment to be developing

the right country do you get the support

for stuff like this I think so yes yeah

yeah generally yes I think UK is a great

place well that's encouraging isn't it

uh on that note um one of the problems

here is not so much the Deep fake news

as the disinformation that is spread by

conspiracy theories who are creating

material they believe to be true what if

we could bring the conspiracy theorist

from the Shadows and back to the light

coming after the break we'll hear about

the AI chatbot that is deprogramming the

people who have disappeared down the

rabbit holes we'll be right back stay

with

us welcome back the moonlandings that

never happened the covid microchip that

was injected into your arm the pizza

pedophile ring in Washington conspiracy

theories abound often with dangerous

consequences many have tried reasoning

with the conspiracy theorists but to no

avail how do you talk to someone so

convinced of what they

believe who is equally suspicious of why

you would even be challenging those

beliefs well researchers have set about

creating a chatbot to do just that it

draws on a vast array of information to

converse with these people using bespoke

fact-based arguments and the debunk bot

as it's known is proving remarkably

successful joining us on Zoom is the

lead lead researcher Dr Thomas Costello

he's the associate professor in

Psychology at the University of

Washington you're very welcome to the

program tell us what the demystified

chatbot

does yeah sure thanks I'm happy to be

here so the the idea is that studying

conspiracy theorists and trying to

debunk them has been pretty hard until

now because there are so many different

conspiracy theories out there in the

world and you need all of all of this

like that you need to look across this

whole Corpus of information uh

comprehensively to debunk all of them

and study them in a systematic way and

large language models these AI tools are

perfect for doing for doing just that um

so so we ran an experiment where we had

people come in and uh describe a

conspiracy theory that they believed in

and felt strongly about uh the AI

summarized it for them and they rated it

and then they entered into a

conversation with this debunk bot uh

where so it was given exactly what they

believed and and programmed set up to to

persuade them away from The Conspiracy

Theory using facts and evidence what we

found at the end of this about 8 minute

conversation this back and forth was

that people uh conspiracy theorists

reduced their their beliefs in their

chosen conspiracy by about 20% on

average and actually one in four people

came out the under the other end of that

conversation actively uncertain towards

their conspiracy so they were newly

skeptical and and so is he the basis

that they don't know where to go to get

this information and they are suspicious

of anybody that might have the answers

to the things that concern them yeah I

mean that that could be part of it I I

think really it's just uh being provided

with facts and information that's

tailored to exactly what they and how do

you how do you deploy it because I don't

I don't I can't imagine that conspiracy

theorists are wandering around saying

disprove the conspiracy theory that I

believe to be true yeah no I mean that's

a great question I think it's one that

uh like I'd be curious to hear others

answers about too in the in the studies

we PID people to come and do it um that

said I'm optimistic about uh you know

the truth motivations of human beings in

general I think people want to know

what's true and so if there's a tool

that that they trust to do that then

then all the better yeah miles can you

see a purpose for this in

America yeah I mean I can certainly see

this principle being incorporated into a

lot of technology I mean a lot of us

already every day use things like chat

GPT and I'll actually give you an

example Christian of chat

GPT disproving something for me so

there's a famous uh Winston Churchill

quote a lie gets halfway around the

world before the truth can get its pants

on no quote better describes the

conversation we're having as how fast

this disinformation spreads well guess

what I put that into chat GPT before I

did a presentation on this subject and

said hold on a second that's actually

not a quote from Winston Church Hill

it's a quote from jonath Jonathan Swift

in the 1700s so AI helped me disprove

that misinformation that's been around

for years so yes I think this is

important and it should be integrated

into these Technologies and Christian is

this where the Two Worlds Collide

because presumably there are conspiracy

theorists who believe something so

fervently that they put out AI generated

material as well so if you can deal if

you can deal with the conspiracy theory

maybe you can stop the prevalence of of

fake material yeah potentially um I must

say though that um so this study was

done in the board atory condition so it

will be very interesting to see whether

these results also translate into the

real world um and then also the um large

language models that were used they were

safety fine-tuned um so that means you

know they were sort of programmed to say

the truth and so on um and so if that

safety fine-tuning is not there you know

um they could be used for something we

call Interactive disinformation so they

could be used to convince people of

things that are not true so that's the

big risk that see here and Thomas I've

got a question for you I'm curious just

about how much having good information

actually changes people's mind and the

example I would give is smoking we've

known for decades smoking is bad for you

everybody agrees we've got all the data

to back it up we put labels on it really

clearly and yet people still smoke and

when you talk to a smoker and try to

persuade them to give it up because you

care about them they will sometimes

really entrench in it's really hard to

break not just because it's addictive

but but because they maybe want to smoke

so I see this parallel perhaps with

conspiracy theories in terms of we have

beliefs and information is not always

enough to change it it's not just about

facts it's about something

else yeah yeah that's a great point I

mean I think that the case of smoking or

or drug other kinds of drug use um we we

know that it's bad for us when we start

doing it um they're fundamentally not

about information whereas whereas

beliefs and particularly conspiracy

beliefs are are often descriptive

they're accounts of what went on in the

world that uh you know Al-Qaeda didn't

uh put together the 911 terrorist

attacks it was the government and and so

dealing with claims about the world um

is something that I think is conducive

to informational persuasion in a way

that that maybe uh like uh nicotine use

is not yeah I mean M we we focus so much

on the legislating it's it's the the the

questions I always ask you how far

behind a congress on that what a state

house is doing about AI legis ation but

but what we've shown tonight is actually

that it's the industry

itself that is that is forcing the

change maybe it's not legislation

because legislation is always one step

behind well well Christian I'm going to

give you an embarrassing admission that

proves that point so I was at dinner

last night with one of the creators of

chat GPT gpt3 one of the earlier

versions she worked for Sam Alman we

were talking about the technology and I

complained to her I said you know I was

teaching a course at University of

Pennsylvania and I got lazy and I was

supposed to come up with a list of 25

books on a subject for my students I

said I'm going to look it up on GPT what

are the best 25 books produced it

emailed it out well guess what my

students emailed me and said all of

those books are fake gpt3 came up with a

bunch of fake books and I said this to

her and she said well yeah and that was

bad and it gave chat gbt a bad

reputation in your mind and that's why

we kept improving the models is we don't

want to serve you up false content

because you won't want to work with this

product and so that may not be

heartening to everyone but certainly

those industry improvements move a lot

faster than legislation because there's

a business imperative to get it right

yeah that indeed is the vested interest

that I see uh for a lot of the online

companies and and of course the AI

companies that are developing this stuff

uh we're out of time uh it flies by

doesn't it just to remind you that all

these episodes are on the AI decoded

playlist on YouTube some good ones on

there as well so have a look at those uh

thank you to Dr schro Dr Costello miles

and of course to Stephanie let's do it

again same time next week thanks for

watching