Ways To Spot Disinformation (on social media)
The Age-Old Problem of “Fake News”
Fake news is not news
Fake news has been part of the conversation since the birth of the Free Press. John Adams did not literally call the Philadelphia Aurora “fake news,” however he was not pleased by the way he was often depicted with “false information”, because his information was more accurate, and did not agree for some in the
group. In the margins of his copy of Condorcet’s Treatise outlines of an Historical view of the progress of the human mind, President John Adams scribbles a cutting note. Writing in the section where the French philosopher predicted that a free press would advance knowledge and create a more informed public, Adams scoffed. “There has been more new error propagated by the press in the last ten years, than in a hundred years before 1798,” he wrote at the time. The change feels shocking, and eerily modern. Were John Adams to have written that sentiment in 2024, and not at the turn of the 19th century, it is easy to imagine that at just 112 characters, he might have Tweeted it instead.
AI generated disinformation poses a threat of misleading voters in 2024.
Computer engineers and tech inclined political scientists have warned for several decades that cheap, and powerful AI tools could soon make it easy to allow anyone to create fake images, videos, and audio that can fool voters… and perhaps possibly sway an election.
The synthetic images that emerged were often crude, unconvincing and costly to produce, especially when other kinds of misinformation were more inexpensive, and easy to spread on social media. The threat posed by AI and so-called deepfakes always seemed like they are a year or two away, however deepfakes have been known to exist as early as 2014 when a scientific paper was published by Cornell University. By 2018, deepfakes were tuned enough to make it much harder to detect. (Deepfakes: How it all began – and where it could lead us (the-decoder.com)
The image below is the same person enhanced by AI:
Source: Nvidia 2018
The image below is the same person enhanced by AI:
Source: Nvidia 2018
Sophisticated generative AI tools can now create cloned human voices and hyper-realistic images, videos, and audio in seconds, at a minimal cost. When strapped to powerful Social Media algorithms, these fakes, and digitally created content, can spread fast and far toward targets with a highly specific audiences, essentially taking campaign dirty tricks to a new low.
The implications for the 2024 campaigns and elections are as large as they are troubling. Generative AI can not only rapidly produce targeted campaign emails, texts, or videos; it also could be used to mislead voters, impersonate candidates and undermine elections on a scale at a speed not yet seen.
Ways to spot disinformation on social media
Fake news is not news, it is a catch-all for fraudulent or politically motivated, and purposeful misinformation that can spread quickly. Especially, with the upcoming U.S. Presidential Election on November 5th, 2024. AI on social media has taken center stage to spread images of political figures with both legit and false messages. Fake X (formerly Twitter), Facebook accounts, etc., using Presidential candidates with stolen images and deep fakes.
Categories of fake news include:
- Clickbait – This uses exaggerated, questionable, or misleading headlines, images or social media descriptions to generate web traffic. These stories are deliberately fabricated to attract readers.
- Propaganda – This spreads information, rumors or ideas to harm an institution, country, group of people, or individual. Typically for political gain.
- Imposter content – This impersonates general news sites that contain made-up stories to deceive readers.
- Biased/Slanted news – This attracts readers to confirm their own biases and beliefs.
- Satire – This creates fake news stories for parody and entertainment.
- State-sponsored news – This operates under government control to create and spread disinformation to residents.
- Misleading headlines – These stories may not be completely false but are distorted with misleading headlines and small snippets displayed in newsfeeds.
1. Use a verified fact-checking site
Fact-checking sites can also help determine if the news is credible or fake.
These sites use independent fact-checkers to review and research the accuracy of the information by checking reputable media sources. They are often part of larger news outlets that identify incorrect facts and statements. Popular fact-checking sites include the following:
- PolitiFact. This Pulitzer Prize-winning site research claims from politicians to check accuracy.
- FactCheck.org. This site from the Annenberg Public Policy Center also checks the accuracy of political claims.
- Snopes. This is one of the oldest and most popular debunking sites on the internet that focuses on news stories, urban legends and memes. The independent fact-checkers cite all sources at the end of the debunking.
- BBC Verify. This site, which checks facts for news stories, is part of the British Broadcasting Corporation.
2. When in doubt, check other reliable sources
Search for other reputable news sites and outlets to see if they are reporting on a story. Check for credible sources cited within the story. Credible, professional news agencies have strict editorial guidelines for fact-checking an article.
3. Read beyond the headline
Think about if the story seems unrealistic, or too good to be true. A credible story has plenty of facts conveyed with expert quotes, official statistics and survey data.
4. Look at the author
Perform a search on the author if you are not familiar with them. Check for credibility, how many followers they have and how long the account has been active.
Review other posts to determine if there are bot behaviors. If the content is retweeted from other accounts and has highly polarized political content, it is likely a fake bot account.
5. Check the source of the information
If this story is from an unknown source, do some research. Examine the web address of the page, and look for strange domains other than “.com”, such as “.infonet” or “.offer”.
Check for any spelling errors of the company name in the URL address. Bad actors may create webpages to mimic professional sites to spread fake news. When in doubt, go to the homepage of the organization, and check for the same information. For example, if a story looks like it is from the U.S. Centers for Disease Control and Prevention, go to the CDC’s secured website, and search for that information to verify that it is legit.
6. Develop a critical mindset
Do not let personal beliefs cloud judgment. Biases can influence how someone responds to an article. Social media platforms suggest stories that match a person’s interests, opinions and browsing habits.
7. Determine if it is a joke
Satirical websites make the story a parody or a joke. Check the website to see if it consistently posts funny stories with as satire. One such site is known for doing this, and is The Onion.
8. Watch for sponsored content
Look at the top results for “sponsored content” or a similar designation. These stories often have catchy photos and appear to link to other news stories. They are ads designed to reach the reader’s emotions. Some of these sites may also take users to malicious sites to install malware.
9. Watch for AI-generated fakes
AI continues to evolve, and at a glance, it can be tricky to tell what is real and what is fake. Advances in AI have created crisper and clearer images, and voice cloning is extremely accurate. However, watch for distortions in areas such as hands, fingers and eyes. These parts of the body typically have irregularities in AI-generated content, such as eyes not blinking properly. Also, watch for voice and facial expressions to not line up properly.
Real examples of a deepfakes
For example, in January, a deepfake audio of President Joe Biden was behind a robocall in which New Hampshire voters were encouraged not to vote.
The technology used to create the deepfake technology in January was from AI voice clone vendor ElevenLabs.
Example Trump: This image, was created by a radio host and his team using AI, is one of dozens of fakes portraying black Trump supporters
Trump supporters target black voters with faked AI images (bbc.com)
‘https://www.bbc.com/news/world-us-canada-68440150 What initially gave it away? Lighting of his face compared to everyone else. Everyone else has the typical glow from professional lighting which will generate a glow. Trump does not have that glow, that would still be there due to the lighting. Trump’s hair also gives it away as it looks fake as his hair is not that perfect. Then there is his “hand”, and Trump is known to have short fingers. Beyond that his fingers are far thicker than in the image. look at the tie and shirt. There are shadowing that does not match and appears enhanced.
Example Biden:
False claims of ‘deepfake’ President Biden go viral (bbc.com)
The image on the left is a deep fake. What gave it away? Lighting. The right image has bright lighting to make Biden to look very pale, and having a “dead eye” face. The photo on the left is a real photo. Bonus indicator for the first image, look at the lighting for the microphones and compare it to the background. The lighting is not equal, especially with Biden’s suit; especially compared to the flags. Just as President John Adams called out the Philadelphia Aurora for their fabricated or slanted news and generalizations, misinformation will always part of the of the process in Government and within the workforce. However, with AI as part of the mix it is important to assess
By Dave Broucek, Trusted Advisor and Cybersecurity
Comments are closed