Barack Obama assembled a team of tech experts, including former Google engineers in Chicago, to leverage social media and digital tools, revolutionizing voter targeting, fundraising and political messaging during his campaigns.

Donald Trump's campaign, with the aid of Cambridge Analytica at Mar-A-Lago, harnesses vast data sets to micro target voters and influence public opinion.

The emergence of technologies during Covid, such as zoom, Google Meet and teams revolutionized remote campaigning, allowing Gen Z to mobilize political movements from anywhere.

Fostering a new era of political engagement.

2024 is the first national election where artificial intelligence is playing a decisive role in who will be the next president and who controls the U.S. Senate and House.

Welcome to the first election in the age of artificial intelligence.

I'm Ali Waldman.

Lafayette College, class of 2026.

And I'm Bridget Coté.

Also class of 2026.

And we'll be co-hosting Lafayette's 2024 election broadcasts.

Artificial intelligence.

Our generation didn't invent it.

We inherited it.

Just like we inherited a democracy that now stands at the edge of uncertainty.

Our generation has its share of rumors and acceleration tests, each with opposing views about AI's impact on society.

Many people are using AI to make more efficient and informed decisions, but others are using it to deceive and manipulate information.

It's hard to know who or what information to trust.

One thing seems certain.

AI is everywhere and changing everything.

Our 2024 broadcast in partnership with PBS 39, is dedicated to trying to understand how AI is affecting this election and our democracy.

We've all been told that we can't believe everything we hear or see on the internet.

That's more true now than ever before.

In this segment, we look at the power of AI and what we can do to combat disinformation.

In this report, we look at how AI, social media, and technology are being used to spread disinformation in the presidential election.

My name is Benjamin White and I'm a sophomore.

Govt law major in 2024.

54% of Americans say they get their news from social media.

Platforms like Facebook, Instagram and TikTok allow users to share content without having to verify whether it's accurate.

In this new frontier, the name of the game is virality.

Clicks, engagement.

And with it comes the ability to create and spread misinformation like never before.

We spoke to Brit Paris, assistant professor of library and information science at Rutgers University, about how AI is being used in this election cycle.

I've seen still images, tons of still images.

One that I looked at recently when it was asked to, comment on was an image of Donald Trump that was posted on Twitter, and he was, you know, surrounded by black women who were his supporters who, were purportedly his supporters, but who, all of whom had too many fingers and no toes.

They had hair coming out of their necks.

They were AI generated and there was, audio voice message, purportedly of Joe Biden's voice telling New Hampshire voters not to vote in a 2023 election.

Voice overs, deepfakes, graphics artificially designed.

Maybe you've once stumbled upon something like this.

Maybe I'll just click politics and become a TikToker.

It seems like, chill vibe that I would enjoy a lot.

Well, this may sound like the voice of Vice President Kamala Harris.

This is an audio we created using an AI software online, and we're not the only ones creating these types of audios.

Whether it's created as absurdist comedy or from the attempt to deceive people.

These types of audios are circulating on the internet.

Listen to this AI voice over a former President Donald Trump that we found online.

I've done more for toilets than any other president.

They finally flushed.

Better now the Democrats.

They ruined toilets.

But we made toilets great.

Again.

Again, these sound bites may sound like a real person, but they're not.

They're generated by artificial intelligence.

But so what?

You might think AI voice overs and deepfakes are something we've been warned about for years.

So why does this matter now?

We spoke with Lafayette Dean and Professor Tim Locke on about his use regarding AI.

So right now, AI is kind of a tool in the toolbox, mostly for people who deal in misinformation and disinformation campaigns.

Probably the example that most people are familiar with is the, generative AI, video of Taylor Swift endorsing Donald Trump.

That was famous.

And so you'll see some of those, but it's being used in a variety of ways.

So if I had become a more prominent part of our lives and we'll continue to grow, then what do we do?

How do we detect it?

Let's hear from Doctor Paris again.

I think a lot of these, AI generated videos still leave you with an impression that something's off.

Something to note is that there are elements of humanity that artificial intelligence may struggle to capture, whether that be breaths the person takes or pauses in a person's speech.

However, with campaigns actively taking part in an AI arms race, the technology will get better.

We don't say this to fear monger, or to imply that deepfakes and AI speeches will make it impossible to determine who really said what, but as the technology grows more sophisticated, so must our ability to identify reality from a lie.

Print media is present in holding our hands through fact checking continues to decline.

So it's up to us to be our own fact checkers and the evolving media landscape.

Undecided voters often determine the outcome in tight elections.

But up until now, pinpointing this group has posed challenges.

Now, artificial intelligence is emerging as an important tool and identifying this crucial electorate.

Campaigns want to get the biggest bang for their buck.

Time and money spent on voters whose minds are made up is wasted.

AI has allowed campaigns to scale the process of identifying undecided voters, to generate voter specific messages, and to deliver that message to the voter with much more efficiency.

Jim Messina, a Democratic strategist working to support the Harris campaign through an allied group, explains that campaigns have utilized an untraditional reliance on third party organizations to reach the 4% of undecided battleground state voters.

Both the Harris and Trump campaign's largest expenses has been advertising, according to the Wall Street Journal.

The Harris campaign will spend about $647 million, while the Trump campaign will spend $273 million prior to Election Day.

Using two user stories, we will examine the ways in which campaigns identify undecided voters, generate content, and deliver this to potential swing voters.

Meet Chai Tory, a college student in Pennsylvania from Texas.

She's an undecided voter and hasn't paid attention to politics until this year.

As Tory is growing on Instagram, she's met with an ad make your vote matter.

Register to vote in Northampton County.

The voter registration ad is a result of it.

Address advertising names, addresses, zip codes and other basic information are used to direct ads.

It's easy to target college students because of the college's IP address and zip code.

Texas has voted Republican her entire life.

Tory decides to register to vote in Pennsylvania, where her vote matters.

She researches topics that are important to her, including reproductive rights.

She utilizes the internet as well as social media.

In the following days, she begins to see more content about the election, including a pro-life post on Instagram.

Social media algorithms work to show the user more of what they want to see.

What a lot of people don't realize is social media companies are able to buy web browser data from third party data brokers.

This allows social media companies to better target their users.

In between posts on her feed.

She gets an advertisement from the Trump campaign highlighting his policy on reproductive rights.

Advertisers in campaigns can purchase the same information as social media companies through data brokers.

Using Secretary's IP address, the campaign can send her ads.

At the same time, her algorithm may continue to post political topics.

That's crazy.

I had no idea how connected my social media was to my internet searches.

It's only going to get more efficient with the increased use of AI.

Let's learn more about our second user story.

Conflicted.

Karl.

Conflicted.

Carlson, African-American union worker in the Lehigh Valley, follows election topics closely.

He has found himself siding with Trump on immigration and the economy.

At the same time, he supports Vice President Harris's stance on reproductive rights in the environment.

Karl's a registered Democrat.

He voted for Barack Obama in 2008.

In 2012, however, he voted for Trump in 2016 and 2020.

The Trump campaign identifies Karl as undecided based on prior election data.

Campaigns can purchase this data from state and local governments.

This data includes contact information, demographic information, and political information such as party affiliation and previous voting.

Karl receives an email from the Trump campaign that outlines his policies.

Three days later, he receives another email from his campaign, which attacks the Biden-Harris administration supporter policy.

Campaigns AI to generate emails allowing them to experiment with content delivery.

I run statistics to understand the click rate and interaction that campaign emails receive.

This allows campaigns to find the most effective advertisements.

Thanks to AI, large analytical teams can be trimmed down significantly.

At the same time, Karl receives emails from the Harris campaign.

It includes her policies, but also a video clip in which she explains her support for unions.

We fight for a future where.

Every worker has the freedom to.

Organize, and.

We will pass the Pro act and end union busting once and for.

All.

Campaigns use predictive modeling to make predictions about individuals by using Karl's personal information and his online activity.

The Harris campaign is able to identify him as a union worker.

They then are able to send him union based content.

One evening, Karl watches both Republican and Democratic leaning news stations on YouTube before the Republican News plays.

He receives an ad from the Harris campaign.

YouTube has identified the election as a topic of interest based on Karl's streaming history and other online data sharing platforms are able to purchase the same information as social media companies to suggest content and advertisements.

Why would the Harris campaign advertise on a Republican leaning show?

It allows her to access conflicted voters like Karl, who may be watching two differently opinionated news outlets.

It also allows her to reach swing voters who may be moderately conservative.

Karl continues to receive political messaging, even for nonpolitical videos.

He begins to realize that he's seeing the same three ads from each campaign.

To identify these handful of ads.

Campaigns utilize AB testing.

Campaigns will select a particular topic, produce two different ads, and see which one performs better.

This process is repeated to help campaigns fine tune their messaging.

AI plays a crucial role in scaling this process.

The fundamental process of identifying undecided voters has changed very little since the 2016 big Data election.

That being said, I has allowed this process to scale greatly.

AI has allowed campaigns to reach these individuals with more precision and efficiency.

It is important to remember that every action one takes online is a new piece of data to be used by these companies, and campaigns to identify target individuals.

Microsoft's contract to purchase power from the decommissioned Three Mile Island nuclear reactor highlights the escalating power demands of AI, and what some see as the controversial use of it and navigating the regulatory process.

Pennsylvania has long been at the forefront of America's energy production, from coal to oil to natural gas.

Now, as artificial intelligence reshapes our world, the Keystone State is once again at the epicenter of a major energy shift.

In 1979, Three Mile Island became another example of America's nuclear fears after a partial meltdown led to five days of panic.

45 years later, this site represents a shift in public opinion, with the recent Microsoft proposal to reopen the Three Mile Island nuclear plant.

My perspective is I was 25 miles away from Three Mile.

Island when it had.

Its accident, and, that's pretty darn close.

Especially when the experts were telling you that, well, you know, very concerned about a radius out to 20 miles.

And, there is something about Three Mile.

Island that will.

Always be a little bit of.

I'm not.

Sure that that that facility should start again.

In a striking shift from previous generations.

Many young people today see nuclear power as a solution to climate change, not a threat to be feared.

I definitely see it heading towards nuclear.

I know there is a fair bit of, like, environmental concerns that I think are fairly overblown.

I think that there's just no energy source that's going to be able to provide as much good in terms of like the energy that it's actually able to produce.

I think like the.

The biggest concern is the waste and, the problematic, situations that are created when you have all this nuclear waste and this trash.

That's created, through these power plants.

I think the concerns should be outweighed by the benefits of nuclear energy.

I know that there are a lot of benefits to nuclear power, but obviously, like in the past, like Chernobyl, like the sort of, like catastrophic events.

That's definitely something that's probably a concern, especially to people like where these nuclear power plants are going to be if they're near neighborhoods or schools or something like that, that can be concerning.

In a twist of irony.

Microsoft plans to use AI to navigate the complex regulatory maze that stands between it and the power it needs.

AI is now being trained on nuclear licensing to speed up the costly and long regulatory process.

This self-reinforcing cycle AI clearing the path for its own expansion poses the question, are we witnessing the first steps towards a future of technological singularity?

We asked assistant professor of computer science at Lafayette College to explain the term, and whether the use of AI for nuclear regulatory approval could be considered an example.

There are some researchers, both historically and in the present day, who research topics related to AI, who, believe that, essentially, once we get an AI that is intelligent enough, that that I will begin to invent other AIS, who will then invent other AIS, who will then invent other AIS, and that this will represent essentially a tipping point for human society.

So I would say the only point where that would really start to concern me would be if, say, an AI system was then also signing off on the regulatory paperwork.

Right.

And how do you envision the role of human experts as AI becomes more involved?

So far, I think what we've seen from a regulatory standpoint is, that actually we do expect that these models should be held accountable.

Is one term that you'll commonly hear to stakeholders, right?

And once tasks that the models are performing get past a certain level of complexity, then it's really experts who are best positioned, to interpret, the output of the language model and to determine whether that actually fits with, what one might hope that the model would produce there.

As AIS appetite for energy grows, the revival of Three Mile Island shows AI tools being deployed in unexpected ways.

While I can help streamline regulatory processes, the real question isn't about machines replacing humans, but how to best balance technological efficiency with human oversight in managing our critical infrastructure.

As the election approaches, sophisticated bots and AI generated disinformation are being used to influence public opinion and cause discontent.

This surge in falsehoods is likely to cause conflict and disruptions in the upcoming presidential race, undermining trust in the democratic process.

Director.

We're about 100 days until the election.

Russia has attacked our democracy in past elections.

Are they seeking to do it in this one?

We assess that the Russian government continues to want to influence, and in various ways interfere with our democracy, with our electoral process.

FBI Director Christopher Wray testified that Russia is working to influence the election.

Nefarious state actors like Russia, China and Iran are using AI and bots to radicalize people.

And so discontent in the United States.

We announced a significant disruption of a, generative, AI enhanced, social media and a bot farm, essentially, of the Russians, that was designed, designed to be an influence operation.

And some of the fake fictitious profiles of, of those bots purported to be U.S. persons.

A bot is an AI account run by a computer algorithm, which are fake accounts programed to mimic a real life person.

AI generated bots are inexpensive and allow campaigns to scale content generation that is impossible to achieve with humans.

This means I can create a flood of content that can reach millions of social media users.

A 2023 study found that bots can post 100 times more content than regular people.

So what does this mean for democracy in the United States?

Full of narratives designed to sow divisiveness, discord, undermine us.

And I'm talking about narratives that they were pushing again, purporting to be Americans, but actually Chinese MPs, officers.

Aegon.

Lockean is a real person, and he seems like a qualified person to talk about national security issues and seems trustworthy with his credentials.

He advocates for certain policies to be taken.

Hello, my name is Aegon Chokin.

I am an active member of the International Association for Intelligence Education, the United States Geospatial Intelligence Foundation, and a proud founding member of the Ocean Foundation.

One of Colossians missions is achieving a diplomatic resolution to Russia's invasion of Ukraine, which means Russia being given territorial concessions.

However, as of today, Russia is a threat, a clear threat to the existence of all mankind.

And Vladimir Putin has nothing to do with it.

The threat discussed below is not related to the sanctions our government has imposed upon Russia, nor is it related to Russia's war with Ukraine.

A de facto confrontation with NATO.

He also wants world governments to stop trying to fight greenhouse gas caused climate change.

This is what is happening right now beneath Siberia.

Continuation of this unchecked process will lead to just one highly predictable outcome a global catastrophe capable of destroying not only Russia, but also the entire world.

The comments on all of his videos have a certain pattern.

They keep saying thank you over and over.

These comments are generated by bots Chiaki and fabricates his credentials and his audience.

It is likely that Russia is involved in assisting people like Chiloquin.

According to the digital media outlet Babble, Russia promotes the same ideas that are wacky and argues.

This is a small example of a larger pattern of bots being used to influence politics globally.

Candidates often get criticized for reversing their position on an issue, but what about voters?

Andreas Polak talks to former Lafayette students about how their views have changed since graduating from college.

My name is Andreas Plexus.

I'm a junior at Lafayette College studying English and film media studies.

And today we're here with Carolyn Ashby.

I graduated in 2008 and was a participant of the first ever broadcast with Lafayette College and PBS in 2006.

I can't believe it's been 16 years since I've been here.

You know, I think for me, certainly what's changed is, is seeing student involvement and, students becoming so passionate about issues and wanting to be involved and learn and voice their opinions about issues going on in the world that are important to them.

And another part of this broadcast we're working on is looking at technology and artificial intelligence.

And 2006, 2008, kind of the beginning of the technology era.

So how have you seen technology change over time, and what was the technology like when you were putting together the broadcast in 2006?

So obviously it was it was way more limited than it is now.

We did have cell phones, but, not all the handy apps that that, that you have now.

And, yeah, I think I think just the fact that students have and everybody, for that matter, has access to information at their fingertips is is a huge change.

I think, you know, back then, for me personally, anything I sort of read in the newspaper or read online, like I sort of, accepted that as the news and it was a credible source.

And I think with so many other, so many, various sources that have popped up, you know, there's so many mediums now.

And so I think the benefit, from the transition from then to now, too, is being a little bit of critical of, like where you get information and, and what's accurate and sort of going through like a fact checking process and really like doing my homework to understand issues so that I can make an informed decision on where I stand with that and not just take it for granted.

Technology.

I think we all I you have sort of a love hate relationship with it.

Me personally, I think it's hard.

I have to really protect how much, power it has in our lives.

I think the last election really showed you the power of technology.

And the marketing and technology.

So I think it'll be interesting to see if, you know, I was if the Democrats really, you know, now, being aware of that, how that affects this year and if they are, you know, I think if they embrace it more as the tool that the Republicans did, more recently.

But I went to school, not everyone had a smartphone.

I didn't have a smartphone in college.

And I don't think a lot of peers that either.

Right.

So even just thinking about your day to day communication, was different.

We didn't email on phone.

So email isn't like in our faces every day.

Which I if you think about it politically and the way campaigns raise money, that means the campaign was at my fingertips literally every moment I got no no apps, no smartphones.

So you know that election in some ways is still a little bit analog.

I guess it was really before maybe the, the dawn of.

I feel like that prevalence of technology in our lives here.

We hear with Edward O'Brien, class of 2016.

Thank you for being here Edward.

Thank you Andrea.

Happy to be here.

And you were a first year when you got involved with the broadcast.

What was it like as a first year in trying to get involved so early on in your college experience?

I knew I had a, I knew I had a higher hill to climb.

And so, you know, I was I was active in the Democratic Party prior to coming to Lafayette.

And so to be able to bring a very different perspective, hopefully was value add to the, to the broadcast and, was an opportunity to, shed a different light.

You know, at that point, social media tied in to a growing, part of people's lives.

And so to be able to do vlogs, was something that was different.

And it was so important to the Lafayette community to always be innovating and doing something different, and showing something in a different way.

And so to be able to be that, that vessel for that opportunity was it was a really awesome experience.

What was your view politically and socially when you were putting together the broadcast, and have you seen that change ever since you graduated from Lafayette?

I've definitely become, more passionate about issues.

I think Lafayette instilled, the drive to lead and to think, critically about issues.

You know, I think the small, discussion based classroom sizes, the fact that, you know, a lot of the, the learning environment was conducive to, you know, participation from everybody and sharing your opinion and your thoughts like, that was kind of the beginning of, of really, you know, being interested in hearing other people's opinions and, and, seeking information, to, to sort of educate myself on, views.

And I think that I've sort of carried through that kind of philosophy that Lafayette instilled in me of, you know, being open minded to hearing other people's opinions and other people's views.

But, you know, I've certainly become a more passionate, politically interested person being, you know, in the last 16 years, I, I cared deeply about our country and, you know, it's future direction.

And, you know, I think, I think that's all started with, with really this broadcast experience and the learning environment in Lafayette.

As a mom.

I think for me, you know, one of the biggest topics that I'm really that's on my radar is gun control.

And I think, you know, when you put on a parent hat, that's something that it just, it's hard to look away from.

I think that you've seen, that's been one, I think, a political issue for me that you've just seen, kind of take off numerous.

Students, professors, departments and alumni have dedicated their time and talents towards the final product being aired tonight.

A lot has changed.

Like I said, I feel like the heat has been turned up in the political environment, substantially since 2008.

2008 was Obama, McCain, and, that was a historical election.

Obama was our first black president.

That was huge for America.

And I think that the discourse around the election in 2020, for now has certainly changed.

But I think that I love to see the students being so involved and engaged, in activism in their causes that they really, truly care about.

Because participation is key.

Getting out to vote is key, and and being engaged is the key to the future.

And that concludes our broadcast on AI in the 2024 election from all of us at Lafayette College and PBS.

39 I'm Ali Waldman and I'm Brigitte Coty.

Thank you for watching.

And don't forget to vote on November 5th.