Tacoma Community College Library: The General Election 2024: AI, mis- and dis-information and the 2024 election

Artificial Intelligence and the U.S. general election

Disinformation and misinformation, exaggerated stories, cherry-picked facts, and attack advertising in times of political high-stakes are nothing new. We as consumers of information and as voters we have always needed to check our own biases and then evaluate information.

In some ways, artificial intelligence doesn't change any of this; but it does make it faster and easier, and creates more sophisticated means, to create and spread images and other forms of information intended to deceive and influence.

As with all information, remember your basic evaluation skills:

Authority: Know who the author is and understand their reputation and affiliations [who they work for, or who sponsors their work]. If you don't know, or you can't find out, that is as big a problem as finding out that the author is disreputable. Legit authors tell you who they are, and are trackable, even if you don't agree with them. (Remember: Just because we don't like a piece of information, doesn't mean it's not accurate and truthful!)

Currency: Know when the information, or when parts of the information, was produced. Sometimes people who create propaganda or disinformation will include information that is out-of-date in order to manipulate a story.

Bias: How much is the information biased toward one ideology or agenda? Does it leave out information in favor of exaggerating a point? What about your bias? How much do you want to believe or not believe something is true, and why? Consider: "How might a person or a group be trying to manipulate me to think in a way that benefits them more than it benefits me?" (Yes, winning an election always benefits the candidate more than it benefits us no matter how much we like them or their policies.) and "How much is this information playing into an identity that I value?" Self-identity is important, but remember that the strength of our identities can be vulnerable to manipulation by those who stand to gain from it. And remember, it's ok to change your mind about something or someone!

Credibility: Does it sound right? Yes, sometimes truth is stranger than fiction, for sure, but question the credibility of an image or a story if it seems to stretch the boundaries of the ordinary. Is it too good or too bad to be true? Maybe. Check sources that are outside of your regular information habits - we tend to choose to read what makes us feel safe and validated. Note: This is also a good place to evaluate whether information is either denigrating or glorifying a person or a group of people, painting them as all bad or all good. People are much more complex than this binary, so in this way, those stories would stretch the credibility of that information.

Below: video: CBS News interview: "How to spot AI and disinformation in 2024 election" (5:34 mins)

TCC Library's information evaluation guide
TCC Library's fake news and dis-information guide
Tips for spotting AI disinformation and progaganda guide
from the Southern Poverty Law Center

Below are some examples of how AI can be used to manipulate voters, and potentially, elections

Foreign interference in US elections?
NPR news: 9/24/24: "U.S. officials say Russia is embracing AI for its election influence efforts"
more...less...
Foreign nations with something to gain from specific powerful politicians being elected, including the President of the United States, use a variety of dis-information tactics to influence voters.AI technologies can better enable and accelerate these efforts to undermine the election process.

Voice simulation

Recordings of simulated candidate voices manipulating voters to act in counter-intuitive ways

For example, before the 2024 New Hampshire primary, a robo-call that simulated then-candidate President Joe Biden's voice targetted Democrat voters to not participate in the state primary vote. This kind of AI deception has the potential of drastically limiting the will of the people in the process of free and open elections.

What to do: First consider ways that this doesn't make any sense (a candidate will always want you to vote). One way to check this information would be to go to the candidate's official website; they would include this kind of information, and any other messages to voters, on their site. While you are there, it's a good idea to use their contact information to let their campaign know that you are receiving information like this.

How Artificial Intelligence Influences Elections, and What We Can Do About It
from Campaign Legal Center: Advancing Democracy Through Law

Targetting vulnerable or historically disenfranchised groups

Robo-calls or other media that delivers a message in a language that the AI (and non-AI) manipulators know would target specific groups.

For example, robo-calls directed at neighborhoods that have high populations of naturalized immigrant citizens, disguised as coming from an Elections office, or are simulated voices of candidates, that tell voters incorrect registration or voting day information, or that they need additional forms of ID or that they need to pay a fee to vote. This has the strong potential to keep eligible voters from voting.

What to do: Always check the voter eligibilty information and the voter registration and voting information from your local county. They will always have accurate information and also may have information about voter manipulation scams as well.

Images and video

Visual media has always been an incredibly powerful way to influence people from political propaganda to advertising.

A common saying in the United States is "a picture is worth a thousand words", and whether AI-generated and disseminated or not, images and video are powerful mediums of influence and deception. With AI, images and video can be produced that are convincingly realistic. Coupled with our general tendency for rapid scrolling of information, rather than lengthy engagement with it, and our very human tendency toward wanting to believe things that align with our biases, AI images and video have a gread potential to convince, manipulate, and deceive.

After sing-songwriter Taylor Swift endorsed the Harris-Walz ticket for U.S. President, a number of AI-generated images were produced to make it appear that Swift and Swift fans were endorsing the Trump-Vance ticket. Many of these images were posted to social media, reaching and potentially influencing hundreds of millions voters.

What to do: The bias and credibility check is powerful. First, if an image or video gives you big, outsized, emotions, whether negative or positive, stop and think about your feeling, and if that image is provoking or satisfying a particular bias you have. For example, in the case above, if you feel especially happy or especially angry that an image suggests that Taylor Swift now seems to be endorsing the Trump-Vance ticket, you might take a step back to verify the information that is generating such a big feeling for you. Credibilty of information: While anyone can change their mind, this quick of a change of heart seems unlikely, so do a quick reality check! In this case, you might check Taylor Swift's official website or social media page to verify their endorsement!

After doing a reality-check, and checking your own bias, other things to look for in images/video:

Odd-looking physical features - maybe the hair, a smile, the nose, or jawline that doesn't quite look like a well-known person. AI can generate little tells like that.
Facial movements in a video that aren't quite what we expect from human movement, including strange blinking patterns.
Shadows in strange places, or no shadows at all.
Objects or body parts that are not proportional in the image.
Repeated patterns or images within the image or video, for example, an individual or a group of individuals that get repeated in a crowd to make the crowd-size seem larger.

Where do our own ethics and responsibility come in to play?

There are many ways that both AI-generated and non-AI generated information can be produced to manipulate, disenfranchise, and scam people. There are also many ways of this information being disseminated through both high and low tech means. Most manipulators and scammers use their own means to spread information but they also rely on us by taking advantage of our human nature to do the rest for them. See the tab: Thinking about your thinking about information".

We are ultimately responsible for the information we create and that we spread, and even the recreational spreading of information that we know to be ridiculous and false, can have big unintended consequences.

As AI technologies and AI users become more sophisticated, the need for ethics and responsibility increases.