Don’t Get Tricked By AI

Artificial Intelligence (AI) generators are getting better and more accurate at making photos, videos, and messages look real. While it can be fun to play with the various AI tools, chat platforms, and content generators, dangers are lurking with those who would seek to harm.

With the rapid spread of AI content comes a rise in misleading deepfakes, or altered images, videos, or audio. According to an article released by the FBI and the American Bankers Association, “[Deepfakes] may depict people you know — including friends and family — or public figures including celebrities, government officials, and law enforcement.”

Fake images and messages, including deepfakes, are increasingly being used to trick people into believing false information or even sharing private personal data with bad actors. As AI-created media is posted and shared more widely online, it’s important to learn how to tell the difference between human-created content and AI-generated content.

We’ve been keeping a close eye on AI developments and new trends, and we want to help keep you safe from deepfakes and bad actors using simulated content.

Pop Quiz

AI can be used to create fake news, scam calls, or even fake videos of politicians, celebrities, or current events. These tricks can spread fast online and confuse people. Being able to pause and ask, “Is this real?” protects you and your loved ones.

Take a look at this “Detect Fakes” quiz from Northwestern University and test your skills on identifying real versus AI-created images. Then, read on to learn how to identify fakes and test again.

Don’t feel bad if you don’t perform well. AI is getting so good that even professionally trained eyes struggle to tell what’s real and what’s fake. The most important takeaway is to fact-check copy, images, and even videos you see online before believing the content is real.

Scams Are Getting Trickier with AI

The biggest risks to consumers right now are scammers using AI to make their schemes more realistic. In 2023, the FTC released a consumer alert about scammers using AI to make family emergency schemes and “grandparent scams” seem legitimate.

Criminals are using AI voice cloning tools to replicate audio pulled from videos and voice recordings posted online to sound like your loved one, often claiming they’re in trouble and need money. Likewise, your own voice or face could be copied from social media and used to trick others into sharing money or information.

AI is also being used in phishing scams sent through emails and texts. These scammers can quickly and easily replicate legitimate-looking organizational messages with fewer spelling mistakes. AI chatbots disguised as investors or romantic partners can carry on conversations to earn your trust.

Protecting Yourself

The general rule of thumb for protecting yourself online has always been to be careful about what you share on social media, including photos, birthdays, and videos with voices, so that this information cannot be used against you or loved ones via deepfakes. Create code words with your loved ones in case an emergency does arise.

Use strong passwords or passphrases and two-factor authentication so AI-driven phishing tools can’t break in as easily. It takes seconds for a strong tool to break through weak passwords, and accounts are hard, or often impossible, to recover without two-factor authentication.

Finally, be suspicious:

• Double-check any urgent call or text asking for money or personal details. Hang up and call the person back directly.
• If an offer sounds too good to be true, it likely is.
• Before you share information or photos on social media, verify the source and fact-check from reliable sources.

Spotting Fake Content

Even professionals have a hard time spotting AI-generated content and deepfakes, and it’s only going to get more difficult as the technology advances. However, there are some basic features you can look for:

• Photos: Look for extra fingers, warped backgrounds, mismatched earrings, unrealistic hair and teeth, unnatural shadows, glare in glasses or windows, reflections in water, and unnaturally perfect skin (real skin will have texture and imperfections).  
• Videos: In addition to all of the features of a still photo, in video, pay attention to lips and voice not quite in sync, flat or unnatural voice, blinking that is either too slow or too rapid, stiff movements, and odd movements in shadows.
• Writing: Notice whether the content sounds too formal or repetitive, or offers only generic advice that doesn’t answer specifics, or if responses to chats are instant and vague.

AI Detectives

When in doubt about whether you’ve encountered an AI-generated image or message, you can turn to AI detection tools online. We recommend verifying against two tools, as they often look for different features to identify AI content.

Photos & videos:

• AI or Not – detects images, music, text, and video
• Attestiv – scans for deepfake and alterations to videos

Writing:

• GPTZero – scans text to see if it’s AI-written
• Copyleaks AI Detector – another text detector
• QuillBot – also has a plagiarism detector

Profiles:

• Reverse image search the profile picture on Google Images or TinEye to see if a profile picture is stolen.

AI: Friend or Foe?

While scammers are getting extremely sophisticated with deepfakes and generated content, AI can still be helpful at work and fun to play with. The goal isn’t to avoid AI, but to be aware of how to spot the tricks.