How do I tell if I’m being catfished by AI?
Being catfished is no fun! It kind of ruins your day and may even bring you harm. And I’m not talking just emotional harm but also physical and financial harm.
In doing a little research here are some of the ways in which you can avoid deception and stay off the catfish line. Start by verifying the identity of the person - and this means doing any or all of the following:
o Copy the image of the person and run it through AI TruthTeller®. AI TruthTeller won’t get every fake image right, but we do know a real photo when we see it. On a recent 300 image test, the app was wrong on only three – that’s 99% correct. So, we’re really only talking about the other half of the equation if you’re worried about false positives. If half of the images are AI and half are real on the GenAI images, and we’re only 50% correct, AI TruthTeller will be right 75% of the time overall and this is way better than a blind guess. And by the way, in small tests we land in the 80-85% accuracy range, correctly identifying AI generated images, so you’re really improving your odds of detection by doing this one thing.
o Do a reverse image search with Google Image search. The photo may be real but if it is found on multiple social media profiles with different names, it could mean trouble. If it comes from a stock photo site (e.g. Getty Images, iStock, etc), the person is not who they say they are.
o Copy and paste any and all text messages and chat sessions from the person into AI TruthTeller. Ask the same question twice to see if they give the exact same answer both times. Ask a question like, “Tell me how you felt the last time you watched a move?” AI has no feelings, no compassion. Pay attention to whether the response is personal (e.g. uses “It was a horror show and I felt scared! or is more objective and distant (e.g. Oh, movies are so much fun!)
o Pay attention to response times. If you’re chatting or texting, AI can respond flawlessly almost instantly. People, well they respond on lunch breaks, in the evening or even while driving (not recommended). The point is, it’s disjointed, lumpy and at times flawed. Generative AI can learn these same behaviors, but this is another clue to watch.
o Ask for the person’s LinkedIn profile and review it for timelines, consistency, conversation and connections. See if they are LinkedIn verified.
o Use the AI TruthTeller Facebook Scam Scan (available only in the AI TruthTeller Windows Chrome Extension). FB Scam Scan assesses and scores over two dozen attributes and activity signals in a Facebook profile. When you start the scan, within moments you will see a Low / Medium / High risk decision. If it’s high, be very, very cautious and careful. Always cross-check with the other points in this list. If it’s Low, the person has likely been online for a long time, has many friends, makes visible posts and may even be Meta verified.
o Beware of persistent excuses to avoid video calls. This is still useful, but be cautious. As Deepfakes get better and better, and computing power increases, this will become increasingly less reliable. We live in an age where AI can write code, engage in real-time conversations, create images, audio and video - all based on input from you and information you have put online or shared in any medium (telephone, online, email, text, social media, etc.)
o Meet the individual in-person, in a public place and bring a friend or two with you. There is nothing quite like looking into someone’s eyes and shaking hands to know they are real.
o Run the other way if the person asks for money. This is how they wave the red warning flag and should cause you to immediately disengage. Online, personal relationships, especially new ones, do not ask for money, unless they mean you harm.
o Listen to your gut and if you are religious or spiritual, pray and listen for that still small voice, even feelings and thoughts that come into your heart and mind. You know more than just your physical senses.
In the end, there is no simple, secret answer to avoid online deception. It requires AI awareness, the use of available tools and most importantly, your own lived experience and discernment. In human relations, humans are superior to AI and always will be. AI may have better recall and organization of facts and figures, but when it comes to human-to-human relationships, there is nothing more real than another living, breathing human.