AI Tools Can Reveal Personal Details From Social Media Posts, Experts Warn
Researchers Simon Lermen and Daniel Paleka warn that modern artificial intelligence tools can identify the real identities behind anonymous social-media accounts by analysing small fragments of publicly shared information. Their study shows that AI systems can combine clues such as hobbies, locations, pet names and personal anecdotes from posts and cross-reference them with other online data to accurately match profiles with real individuals. The researchers say this capability raises serious privacy and security concerns, as governments could use it to track anonymous activists while cybercriminals could exploit it for targeted scams like spear-phishing. Experts advise users to limit sharing personal details online and urge social-media platforms to introduce stronger safeguards against automated data scraping and mass information extraction.
Artificial-intelligence tools are making it significantly easier for hackers to uncover the real identities of people behind anonymous social-media accounts, with researchers warning that even casual online posts could reveal sensitive personal information such as where someone walks their dog.
A recent study by researchers Simon Lermen and Daniel Paleka shows that modern AI systems can analyse small fragments of publicly shared information and link them to real-world identities with a high degree of accuracy.
How AI can reveal personal identities
In their experiments, the researchers fed anonymous online profiles into an AI system and asked it to analyse available details across posts. The model was able to cross-reference clues including location references, hobbies and personal anecdotes with other publicly available information on the internet.
In one hypothetical example used in the research, an anonymous user casually mentioned walking a dog named “Biscuit” in a specific park while also referring to struggles with schoolwork. By combining the pet’s unusual name, the location and contextual details, the AI system was able to identify the likely real-world individual behind the account.
Researchers say such techniques highlight how AI can piece together seemingly harmless fragments of data that people share across different posts and platforms.
Risks of AI-assisted profiling
The findings raise concerns about how AI-powered profiling tools could be misused. According to the researchers, the technology could potentially be deployed by governments to identify anonymous activists or by cybercriminals to carry out targeted scams such as spear-phishing attacks.
Spear-phishing involves attackers impersonating trusted contacts or organisations to trick victims into revealing confidential information or clicking malicious links.
The researchers also warned that these methods are becoming increasingly accessible. With publicly available language models and online data sources, individuals with minimal technical expertise could conduct such identity-matching attacks.
How users can protect themselves
Experts say users can reduce the risk of being identified by limiting the amount of personal information they share online. Even minor details such as location tags, routine activities or unique personal references can collectively reveal more than intended when analysed by AI systems.
Researchers also suggest that social-media platforms strengthen safeguards by introducing stricter limits on automated data scraping, monitoring large-scale data extraction and restricting bulk exports of user information.
Until such protections are implemented more widely, experts advise users to remain cautious about posting location details or other identifying information, noting that AI tools can now combine scattered clues across the internet to build detailed personal profiles.
References
- Using AI, Hackers Can Even Tell Where You Walk Your Dog, Here's How To Prevent That