A recent study has highlighted how artificial intelligence can be used to pinpoint locations from social media posts. Conducted by McAfee, the research indicates that various AI tools are capable of detecting the location of images shared online, even when GPS information or explicit location tags are absent. The findings suggest that, on average, AI tools could accurately identify the correct city and country in approximately 90 per cent of instances.

The AI models evaluated during the study included Google’s Gemma 3 27B and Alibaba’s Qwen 3 VL 30B. Both models are accessible for public use without any subscription and demonstrated notable effectiveness in location identification.

Mechanics of AI Location Detection

The methodology employed by the researchers involved providing the AI models with travel images devoid of any accompanying metadata, EXIF data, or file names. The AI was prompted to ascertain the location based solely on the visual details present in the images. Human reviewers further analysed cases where the names of places could yield multiple interpretations, ensuring reliability in responses.

Among the 21,236 travel photographs analysed, Gemma 3 27B correctly identified the city and country in 87 per cent of cases, while Qwen 3 VL 30B demonstrated a slightly higher accuracy rate of 91 per cent. The AI analyses factors such as architectural styles, signage, distinctive skylines, lighting conditions, street marks, storefront characteristics, and even food vendors to arrive at its conclusions.

The study noted that images from recognised tourist destinations exhibit clearer identifiable features, making them easier for the AI to classify. Conversely, photographs from generic beaches, rural roads, and indoor hotel settings proved more challenging for the AI models, although the country of origin was often still identified successfully.

Implications for Cybersecurity and Scamming

This research raises concerns regarding the potential for scammers to exploit AI models for enhanced fraud tactics. Reportedly, individuals could manipulate public travel photography from platforms such as Instagram, Facebook, or X, utilising AI to glean likely destinations and timeframes. Armed with this information, scammers might customise messages, increasing the likelihood of success.

For example, fraudulent warnings could be sent to individuals, claiming unusual banking activities during their travels or suggesting attempted logins from their location. The precision of these messages is less critical than their plausibility; they need only be credible enough to evoke concern.

With a notable surge in cybercrime reported recently, as cited by the Ministry of Home Affairs, which recorded 28.15 lakh cases in 2025—an increase of 24 per cent from the previous year—the potential for AI to facilitate such scams becomes alarming. In 2024, cybercrimes resulted in financial losses of Rs 22,495 crore for Indians.

Users are advised to exercise caution when sharing images on social platforms, particularly if the photos lack proximity to well-known landmarks. The capacity of AI tools to accurately determine specific locations diminishes without distinctive visual markers, but the risk associated with oversharing remains significant.

Long or Short, get news the way you like. No ads. No redirections. Download Newspin and Stay Alert, The CSR Journal Mobile app, for fast, crisp, clean updates!

App Store –  https://apps.apple.com/in/app/newspin/id6746449540 

Google Play Store – https://play.google.com/store/apps/details?id=com.inventifweb.newspin&pcampaignid=web_share