Researchers have trained computers to spot social media users who pose as somebody else – a practice known as catfishing.
They say their algorithms can identify users who lie about their gender, with more than 90% accuracy.
Most dating website users say they have encountered at least one fake profile, according to consumer group Which?
And the number of people defrauded by dating scams reached a record high in 2016.
Analysing data from 5,000 verified public profiles manually checked by employees on adult content site Pornhub, the algorithms learned how men and women of different ages interacted with others, how they commented on posts and their style of writing.
The study suggested almost 40% of the site’s users lied about their age and 25% lied about their gender, with women more likely to deceive than men.
Dr Walid Magdy, of the University of Edinburgh’s School of Informatics, said: “Adult websites are populated by users who claim to be other than who they are, so these are a perfect testing ground for techniques that identify catfishes.”
“What was interesting was that it seems that for many the reason for lying was to get more friends and subscribers.”
Dr Magdy said the algorithms, developed by computer scientists at Edinburgh University, in collaboration with Lancaster University, Queen Mary University, London and King’s College, London, could “lead to useful tools to flag dishonest users and keep social networks of all kinds safe”.
“It has many applications such as people who fake accounts on Twitter for political reasons or for children who fake accounts to access adult websites,” he said.
The study will be presented at a conference in Australia on the future of social networks.