Tech & Science
…
…
Tech & Science
Herz — Tech & Science Desk · · 30s summary · 2 min read
Scammers can use artificial intelligence to analyze vacation photos posted on Instagram or Facebook and determine travelers' locations, making their fraudulent messages more credible. McAfee tested over 21,000 images with two free AI models, which correctly located 91% and 87% of them. The analysis can work without declared geolocation or file metadata. Waiting to post until after returning home and limiting visibility to known contacts reduces information accessible to fraudsters.
Scammers can analyze vacation photos posted on Instagram or Facebook using artificial intelligence. The location they deduce can lend credibility to fraudulent requests for banking information, according to The Guardian.
McAfee submitted over 21,000 travel images to two free AI models. These models correctly located 91% and 87% of the images, respectively.
The analysis does not necessarily depend on declared geolocation or file metadata. Background details, architecture, signs, and lighting also provide clues.
Recognizable landmarks, city skylines, road markings, food stalls, and storefronts also facilitate location identification by AI.
Scenes containing distinctive elements are easiest to locate. Beach and hotel room photos yield less precise results, but the country can often still be identified.
In one test, ChatGPT correctly located a photo showing a river and trees in the foreground to Hastings-on-Hudson, New York State.
A photograph of flowers was also correctly linked to Keukenhof Gardens in the Netherlands. The arrangement of tulips and small blue flowers provided the clue.
The identified location allows scammers to personalize fraudulent messages. A scammer can mention unusual activity on a card, a specific hotel, or a login attempt from a given country.
These details can match the vacation visible on social media and strengthen the credibility of messages aimed at obtaining banking information.
It is advisable to wait until returning home before posting vacation photos. Their visibility can also be limited to known contacts.
The illustrations in this article are generated by artificial intelligence.
No comments yet. Be the first to react.
Any message demanding immediate action should raise suspicion. It is recommended not to click on links in such messages.
To verify an alert, contact the company or bank directly. Credentials should come from their official website or the back of your bank card.
Yes. AI can exploit the background, architecture, signs, road markings, and lighting.
Photos containing recognizable landmarks, city skylines, signs, food stalls, or storefronts provide more clues.
They can mention unusual card activity, a specific hotel, or a login attempt from the relevant country.
It is advisable to wait until returning home and restrict visibility to known contacts.
Do not click on any link. Instead, contact your bank using the number on your card or from its official website.