Social Media Fact Check: How to Avoid Spreading Viral Misinformation and Detect AI Images – by Carly Winchell

These two images were generated by the AI ​​within minutes with the prompts “Kamala wearing a Trump hat” and “Trump with kittens.” Note the nonsensical labeling on the hat in the first image and the surrealist airbrush quality of the second. Two telltale signs of AI-generated images.

2024 is a year in which misinformation spread online gained national attention through the presidential debate. Fact-checking information found online, particularly on social media sites, is an important skill for both politicians and voters.

But in a world of artificial intelligence (AI)-generated media and intentional disinformation, what steps can people take to avoid contributing to this misinformation problem?

Avoid sharing posts reflexively.

The first step to avoiding the spread of misinformation is to pause and resist the temptation to share right away.

Misinformation is often presented in an inflammatory manner to upset or upset people. In this emotional state, people may want to share a post immediately because it seems urgent to them; However, sharing a post that contains misinformation can help it spread and go “viral.”

Even if the user later realizes that the post contained incorrect information and corrects themselves, the damage has already been done. The post has probably already been seen and shared by many others.

Fact-check the post before hitting the share button.

This Star Wars: Solo model features Adam Driver playing the titular Han Solo. The image was accompanied by text saying that a reboot of the film was in the works. A quick search on the Internet reveals that this information is incorrect. Screenshot from Facebook.

If you come across a post with questionable information, it's important to check the facts. If some things seem too good to be true or too outrageously bad to be believable, they are probably false. Therefore, it is important to check the content of the information.

Fact-checking websites are available to assist with this process (some linked at the end of this article). However, a simple first step is to search the internet for reliable evidence to support the claim.

“Reliable” is the most important word. Using other social media posts, blog opinion pieces, or other potentially unreliable sources can only amplify the misinformation. Instead, look for reputable sources.

Depending on the content, there may be a variety of credible sources from which to verify the facts. If a claim is made about a particular topic, look for sources that have reliable information on the topic.

For example, a meme might claim that violent crime rates in the United States increased in 2023. A social media site is not the most reliable source of crime statistics. Therefore, before sharing the meme or information with others, it is advisable to do some research to verify it.

A reputable source of the country's crime could be the Federal Bureau of Investigation (FBI), which collects crime statistics.

On September 23, 2024, the FBI issued a press release indicating that overall national violent crime decreased 3 percent in 2023 and murder and manslaughter decreased 11.6 percent compared to 2022.

Since the information in the meme is not true, it is now clear that the post should not be shared as it would spread misinformation.

When determining the credibility of a source, the following points should be considered:

• Currency: When was it released? Every year new information is discovered. Information about computers in 1984 will be very different from that in 2024. The FBI violent crime release is very current and is for the correct year, 2023.
• Author: Who wrote or published it? Is there a motivation for bias? A politician who wants to win votes has more motivation to mislead than a researcher who publishes an article in a scientific journal. PETA (People for the Ethical Treatment of Animals) will have different motivations than the Cattlemen's Association. The FBI tracks national crime statistics because it is a national law enforcement agency.
• References: Does this author have the credentials to support his claims? Nobody knows everything about everything. An expert in his field has more reliable information than an opinion blogger. A meteorologist knows more about the weather than a heart surgeon. The FBI will know more about crime statistics than Uncle Steve.

Artificial intelligence (AI) and deepfakes

At first glance, this AI-generated image of Kamala Harris may seem fine, but a closer look at the details reveals inconsistencies. One of her fingers is pointed backwards and a kitten's head is coming out of Harris' face.

When AI is involved, the water becomes even murkier. AI-generated images and videos can be credible at first glance and sometimes even stand up to scrutiny.

The first step to identifying AI images is to look closely at the details. AI tends to have problems with finer details.

Human hands may look strange or unnatural, or the text in the image may be jumbled or nonsensical. The background can also appear imaginative or inconsistent.

Sometimes these clues are missing or difficult to recognize. Viewers should also be wary of images that look too perfect to be true (which is often the case).

Another trick is to do a reverse image search. AI images tend to appear less frequently than real ones, or a reputable source may have already fact-checked the image.

A fake AI image of Trump wading through the floodwaters was shared over 150,000 times in 16 hours. Image courtesy of Snopes.

An article from fact-checking website Snopes breaks down a recent AI image of Trump wading through floodwaters to help storm victims. The image was created by AI but shared on social media as if it were true, accompanied by the false claim that Facebook was censoring the image to harm Trump.

The post with the fake image was shared more than 150,000 times within 16 hours.

Always ask questions.

It's important to approach memes and other information on social media with a healthy dose of skepticism.

People may inadvertently share misinformation in an attempt to help others by shedding light on a concerning issue. However, those who want to spread false information for their own purposes rely on this motivation to get people to spread lies without questioning them.

Stopping, researching, and sharing only accurate information can help stop the spread of viral misinformation.

Practical application and resources

Consider the following meme shared on a local Buena Vista Facebook page:

Gasoline price memes. Screenshot from Buena Vista Water Cooler Facebook page.

This meme suggests that the economy is doing worse under the Biden administration because gas prices have increased. But what does the development of gas prices actually look like?

From 2016 to 2020, retail gas prices fluctuated between 1,764 (February 2016) and 2,836 (September 2018), according to the US Energy Information Administration.

From 2020 to 2024, it fluctuated between 1.841 (April 2020) and 4.929 (June 2022). At the most extreme differences (February 2016 and June 2020), this meme may seem true, but these statistics are outliers.

Looking at a full year during each term (2019 and 2023), the difference becomes smaller at 2.604 compared to 3.519 (a difference of 0.915).

The truth is more complicated than portrayed in the meme. The author has taken the most extreme data points and exaggerated them to make a political statement.

Additionally, it is important to remember that gas prices are not the only measure of the health of an economy.

Scott Fulford, senior economist at the Consumer Financial Protection Bureau, said in an article published by Princeton University Press: “Employment, the broadest measure of economic success, was higher in 2023 than in 2019. In 2023, it was as if the pandemic “Hadn't done it.” is what happened from a macroeconomic perspective.”

Fact-checking resources:

Screenshot of recent fact-checked statements from Politifact.com.

Below is a list of resources that can be used to fact-check information appearing on social media this election cycle:

• Primary Sources: Focuses specifically on verifying viral stories.
• Media Bias/Fact Check: Used to determine the reliability of media resources such as Fox News, CNN, MSNBC, etc.
• Snopes: Fact-checks rumors, videos, photos, claims, etc. Accepts fact-checking submissions.
• FactCheck.org: A project of the Annenberg Public Policy Center that monitors the factual accuracy of statements made by political actors in the USA.
• AP Fact Check and Reuters Fact Check: Address misinformation surrounding breaking news.
• AFP Fact Check: A department of Agence France-Presse that deals with misinformation internationally.
• Politifact: The nonprofit Poynter Institute for Media Studies rates politicians' statements on the Truth-O-Meter, ranging from “true” to “pants on fire.”