Fake ID’s generated by LLM AI programs
I have always tried to warn hosts about getting a false sense of security from requesting guests send a picture of their ID when renting. Our policy is to only check them at check in. Then we are able to check for the required security features such as holograms and under-printing, and we can handle the card to see if it is solid or some badly laminated dupe. Then we also have the ability to swipe the ID into our computer system (like a hotel would) which pulls relevant data as allowed by law in my country, which also would flag anything that wasn’t authentic.
Technology is rapidly outpacing your ability to verify authenticity when you have contactless check in’s. How you handled that is up to you, but I wanted people to be aware of a trend that I have been warning about here for over a year.
“Some images were more convincing than others. The fake medical prescriptions were legible, but the handwriting looked more like the output of an iPad stylus than a pen on paper. When I fed OpenAI’s model a boarding pass from an old flight and asked the bot to update it with new details for an upcoming flight, ChatGPT generated a new boarding pass—but surely, the bar code wouldn’t have actually scanned me onto a flight. And although I certainly hope my ChatGPT-generated driver’s license would not fool the TSA, perhaps it would trick a hotel receptionist or an out-of-state bouncer who would accept a “photo” of my ID instead of the real card. “
https://www.theatlantic.com/technology/2026/05/chatgpt-images-deepfakes-fraud/687023/