April Fools’ Day has long been a time for pranks, and harmless tricks. But while most April 1st jokes are meant to amuse, history has shown that tricksters aren’t always so innocent. From ancient con artists to legendary hoaxes, deception has been part of human nature for centuries. Fast forward to today, with the help of technology, scammers have taken the art of deception to a whole new level.
Artificial intelligence (AI) is revolutionizing how scams are carried out, making them more convincing, harder to detect, and more dangerous than ever before. Even though AI services like ChatGPT have safeguards in place to prevent malicious use of their software, creative prompts can get around such restrictions. Additionally, adversaries have created their own unsupervised learning systems that lack the guardrails put in place by ChatGPT. One of these, FraudGPT, can be found on the dark web and can be used to craft highly convincing phishing emails among other things.
Let’s look at three common ways that malicious actors use AI for deception:
Deepfakes
Deepfake technology uses AI to create realistic fake videos and images. Scammers can manipulate videos to impersonate public figures to get you to spend money or influence your opinions. They can also impersonate people you know. In one case, a finance worker paid out $25 million dollars to scammers after a deepfake video call with his “CFO”.
Voice Cloning
AI can now replicate someone’s voice with just 30 seconds of audio. Imagine someone calls you to ask if you want to buy an extended warranty for your car, and when you decline, they ask you a few more short questions designed to capture more of your voice. You may not realize it, but your voice has been stolen forever, and a scammer can now make you say anything they want. Scammers can use this technique to impersonate people you trust. Because the voice sounds exactly like your child, spouse, or boss, your instinct is to react immediately and not think before you act. In just a few moments, you’ve fallen victim to the scam by wiring money or sharing sensitive data.
Phishing Emails & Messages
Phishing scams use AI to craft highly convincing emails and messages that appear to come from legitimate sources. In the past, spotting phishing emails was easier because they often contained poor grammar, spelling mistakes, or awkward phrasing. AI can now generate flawless and highly personalized text, making phishing harder to detect.
How to Protect Yourself
While there may currently be ways to identify subtle abnormalities in AI-generated images and text, AI technology is rapidly advancing, and these techniques may soon become unreliable. Rather than focusing on whether content is real or AI-generated, the key concern should be whether it is being used to deceive you in some way.
The most important question to ask yourself when encountering any communication is this:
Is this communication unexpected, and does it ask me to do something I have not done before for this requester?
If the answer is yes, STOP and THINK before acting. Always verify the request through a separate, trusted method before responding or making decisions.
In the same way that April Fools’ Day reminds us not to believe everything we see or hear, today’s digital world requires a heightened sense of skepticism. The difference? AI-powered scams aren’t just pranks—they’re serious threats that can cost people their money, privacy, and security.
