Increasing sophistication of fake posts is making it harder to identify them, but there are ways
Posted
Lee Shappell
About Civility Learning Lab
In 2025, Independent Newsmedia, USA, Inc. launched Civility Learning Lab, a free, non-subscription platform designed to promote civil discussion of issues affecting Arizonans. It now includes INI's Delaware/Maryland and Florida newsrooms. CivilityLearningLab.org is made possible by sponsorships from INI and retired INI CEO Joe Smyth. For more information on the platform or how to become a sponsor, email CEO Charlene Bisson at cbisson@iniusa.org.
OPINION — Early this summer, an “advertisement” about a big sale — $1 books at Barnes & Noble — was all over social media. Any voracious reader would hoist an eyebrow.
Last spring, there were widespread “offers” on the internet of deeply discounted tickets and travel packages to see the University of Arizona men’s basketball team play at the Final Four. Great for starving students or people who love the Wildcats but are fiscally challenged in the age of shrinking affordability.
These two examples had little in common other than they were both fakes.
The internet is filled with them. Many are far more subtle than these too-good-to-be-true examples. Scammers want your personal information — and likely your money. They’re counting on you to be naïve to their ways.
They titillate you. They appeal to your quick-reflex reaction to part with your credit card and other personal information. TransUnion’s Top Fraud Trends Report, released in April 2026, found that 1 in 6 U.S. consumers lost money to digital fraud in the past year.
Artificial intelligence-driven scams are spreading fast and getting better. AI tools have made it easy for scammers to build convincing copies of legitimate brand websites in minutes. They no longer need advanced coding or design skills. They can paste a real brand’s URL into an AI tool to duplicate layouts, logos, color schemes and text instantly.
They often use phishing, with fake emails, text messages or websites that look real, to steal passwords, credit card numbers or money.
Deepfakes — images, videos or audio recordings created or altered by AI to make a real person appear to say or do something they never actually did — are AI-powered studies of thousands of real photos, videos or voice clips of people to learn their exact facial expressions or voice. They can make almost anything look or sound real.
So, how do you distinguish fact from fiction? There are tools and techniques, but often it comes down to trusting your own good sense: If it sounds too good to be true, it probably is.
Pump the brakes
The best advice: Not so fast. Do some research. It’s fast and relatively easy, and it may prevent you from unknowingly coughing up the keys to your personal vault.
As fakes powered by AI get better, tools powered by AI also are evolving to spot them.
First, know some of the “tells” of a fake.
Phishing scams try to scare you with sudden account blocks. They usually have odd links, with web addresses that look strange or are slightly wrong. Impersonators often add extra letters or the word "official" to a web address to make it look real.
Their greetings might not have your real name. They might ask for passwords or money.
If there’s no means to contact the sender or they do not provide a physical address, beware.
Do they provide detailed shipping, return or privacy policies? If not, that’s another red flag.
Speakers’ or writers’ cadence and rhythm may contain sentences with no variety in pace or emphasis. Their tone may be of questionable propriety for the subject, too flat or too stilted. That’s not right.
Videos or photos may have small details that give them away. They might look too smooth or too perfect. Check details like hands. Are there extra fingers, or do fingers blend together like melting wax?
Enlarge the photo and look carefully at faces. Smooth, or are there pores and imperfections that real people have? Are pupils of differing sizes or shapes? Do the “people” in videos ever blink?
Do joints bend in weird ways? Does text on shirts change as a video rolls on? Do things like fences, walls or doorframes curve? Do objects seem to hang in the air with no support?
Do shadows point in different directions or not match the light source?
Another red flag: AI struggles to copy company logos and often spells words incorrectly in the background. That screams “fake.”
What you can do
The best place to start is straight from the horse’s mouth: Go to the source, the official site. In the Barnes & Noble scam, a quick visit would have exposed the $1 book “ad.”
“This is a fake online promotion and most likely a phishing scam,” the company said in a statement in July. “We encourage customers to go to our website to see current legitimate sales and promotions.”
UofA hoops fans could have visited the Wildcats’ ticket office to see that the ticket/travel “offer” was fake.
Do not click strange links from pop-up ads or unverified social-media posts. Look online to see if an ad or offer has been reported as a scam.
In seeking the truth, adopt a “says who?” attitude. On what site did you see the claim? Have you ever heard of the site? Is it the official site? Do a search. Are other sites also reporting the information? If the website where you saw the information is a stranger to you and no one else has that information, be wary.
Sites like Infosec show you how to be a scam detective.
Tools like Google Safe Browsing can help you verify if a website is dangerous or fake.
Guardio is among the browser-based cybersecurity extension and mobile apps that protect users from phishing, malicious websites and online scams in real time. It is a subscription service that essentially uses AI to fight AI, to identify scam shopping sites that look 100% real.
The News Literacy Project, founded in 2008 by a Pulitzer Prize-winning investigative reporter and used widely by schools and educators to teach students how to spot misinformation and evaluate credible sources, has a series on AI with free educator tools for teaching about artificial intelligence and the critical-thinking skills needed to navigate and use the technology.
Rumor Guard, operated by The News Literacy Project, evaluates viral rumors and misinformation using clear, transparent fact-checking standards.
At Entrust, explore the latest in global-fraud intelligence.
MarqVision, for businesses, is another protection that uses AI to combat scammers who use artificial intelligence to copy brand websites and trick shoppers into buying counterfeit products — or steal their personal information.
For photos or videos, Hive Moderation AI Detector offers a free web demo where you can upload images or drop URLs to test for synthetic content. It is widely used for high accuracy on modern image models.
AI or Not is a fast platform designed for quick, casual checks of images and videos. Upload the file and receive a percentage-based verdict.
TruthScan is a specialized checker built to recognize pixel-level patterns from newer generative models like Flux and Google DeepMind outputs.
Finally, a great way to protect yourself is by using strong, unique passwords as well as enabling two-factor authentication.
As generative artificial intelligence — which creates text, images, audio, video and code based on patterns it learns from existing data — permeates our lives, it has never been more important to understand the technology and its potential.
Just bear in mind that information today to combat fakes might be of little to no use a year from now as the technology morphs at light speed. But as fakes powered by AI get better, tools powered by AI to spot them also are evolving.
Lee Shappell is opinions associate with Independent Newsmedia, Inc., USA’s Arizona newsroom. Please submit comments at yourvalley.net/letters or email them to AzOpinions@iniusa.org. We are committed to publishing a wide variety of reader opinions, as long as they meet our Civility Guidelines.
Comments
No comments on this item Please log in to comment by clicking here