A recent investigation revealed that writing online reviews can make a person more vulnerable to phishing attacks. The study, published in May in the journal Information Systems Research, demonstrated that seemingly insignificant behavioral data, such as the length of reviews written by a user, can be used to infer hidden social connections with considerable accuracy.
The researchers created a model called Homophilous Network Learning (HNL), which analyzes specific patterns present in online reviews. This model examines elements such as the establishments reviewed by users and the number of words used in their reviews to identify possible social connections between people.
To test the model's effectiveness, the researchers used public data from the Yelp platform, a U.S. commercial establishment review service. The choice of this platform was due to the availability of public data that allowed testing the HNL model's capabilities in identifying hidden social networks.
The study results warn of the privacy risks associated with sharing information online. Although reviews may seem like simple, trivial actions, cybercriminals can exploit this seemingly harmless data to obtain personal information and create more sophisticated and targeted phishing attacks.
