‘We detected unusual activity’: the scam that uses AI to exploit your holiday photos

Fraudsters are exploiting holiday photos posted on Instagram and Facebook by using artificial intelligence (AI) to extract personal information and trick victims into revealing their bank account details. After travelers post pictures of their stay, AI can analyze the background, architecture, signage, and light to pinpoint the location of the trip.

This added credibility to the scam increases its chances of success. McAfee, an anti-virus software company, tested two AI models on over 21,000 travel images and found that one model identified 91% of the images accurately, while the other achieved an 87% success rate. When asked to replicate the experiment, human staff members found it easy to identify locations from travel photos.

The AI's success rate increases when the photos feature recognizable landmarks, skylines, signage, street markings, food stalls, storefronts, beaches, or hotel rooms. However, the accuracy drops when the pictures are of a beach or a hotel room. If a scammer intercepts a photo, they can use the information to send convincing messages, such as claiming that the victim's card was compromised due to unusual activity during their travel or asking to confirm their identity due to suspicious login attempts from a specific country.

To avoid falling for this scam, it is recommended to postpone posting holiday pictures until after returning home and adjusting privacy settings to limit visibility. Additionally, people should be cautious of urgent messages requesting immediate action and should never click on links provided in texts or emails. Instead, they should contact the company or bank directly using the information from their website or bank card.