Fraudsters use pictures posted on Instagram or Facebook to create emails seeking bank account details<\/p>\n
You are on a short break in Porto and post some pictures of your family on Instagram or Facebook. With a small section of the Douro river in the background, you think it could have been taken anywhere.<\/p>\n
A few days later, you get a text saying that your card was compromised. \u201cWe detected unusual activity while you were travelling in Porto \u2013 please verify immediately,\u201d says the message.<\/p>\n
You click on the link to confirm your bank details. Since you have not put any details of your trip on any of your social networks \u2013 bar the indistinct pictures \u2013 you don\u2019t suspect there is anything suspicious about the messages.<\/p>\n
But the text was a fraud designed to extract your financial details. The criminals behind it figured out where you had been on holidays \u2013 lending credibility to their text \u2013 by using AI to analyse the image for the most sparse signs of where it was taken.<\/p>\n
New research has shown that by picking up on details in the picture \u2013 such as the background, the architecture, the signage or the light \u2013 the AI agent can pinpoint where it was taken. For criminals, this information gives attempts to defraud by text message or email an added legitimacy.<\/p>\n
McAfee, the producer of anti-virus software, used two freely available AI models to test more than 21,000 travel images. One of the models identified 91% of images accurately while the other got 87%. Staff were then asked to replicate the experiment with their own pictures and became uncomfortable with how easy their travels were pinpointed.<\/p>\n
The company says that it shows that computers don\u2019t need photos to be tagged or attached metadata to identify where they are taken.<\/p>\n
\u201cWhat AI does is give context \u2026 so that makes the scam [and] makes the threats credible,\u201d says Vonny Gamot, the head of EMEA at McAfee.<\/p>\n
Pictures are more likely to be identified by AI if they have recognisable landmarks, skylines, signage and street markings. Food stalls and storefronts can also pinpoint quickly where the picture was taken.<\/p>\n
If the picture is taken on a beach or a hotel room, the accuracy of the AI is lowered. But McAfee reports that it is likely that the system can identify which country they were taken in \u2013 which is all the scammers need.<\/p>\n
When a staff member tested ChatGPT to identify a picture of a river with some trees in the foreground, it correctly pinpointed it as Hastings-on-Hudson, an area in New York state.<\/p>\n
Another picture of a group of flowers was identified as the Keukenhof gardens in the Netherlands. The AI agent correctly deduced that the layout of the tulips, along with smaller blue flowers planted between them, meant that it was the famous gardens in the picture.<\/p>\n
Criminals can use the information to make their approaches more convincing. They might say that your card was flagged for unusual activity while you were somewhere. Or that they are calling you after your stay in a particular hotel. Or that they want to confirm your identity because there has been an attempt to log in to your account from that country.<\/p>\n
If you want to post pictures, delay until after you get home and change your settings so that only the people you know can see them.<\/p>\n