Researchers have developed an artificial intelligence system capable of identifying fake profiles on dating apps with 93% accuracy, according to a new study. The neural network analyzed profile data including age, gender, ethnicity, and the language used in self-descriptions to distinguish genuine users from scammers.
The collaboration between computer scientists in the UK, US, and Australia trained the system using nearly 15,000 profiles from the free dating site Dating 'N More, which publicly posts detected fake profiles. The analysis revealed that 60% of fake profiles belonged to men, with an average age of 50.
Fake profiles were found to use more images and emotive language, commonly employing words such as 'caring', 'passionate', and 'loving'. The researchers aim to develop an early warning system that can flag scammers as they create accounts and begin contacting potential victims.
Statistics from Action Fraud, the UK's national fraud reporting centre, indicate that British victims lost £50 million to romance scams in 2018, with 63% of victims being women. The researchers note that current countermeasures, such as blacklists, are easily circumvented, whereas their AI-based approach could prove more robust.
