A nice catfish smiling and swimming in water

Neural network taught to recognise fake PornHub profiles

Image credit: Dreamstime

Computer scientists at the University of Edinburgh have developed a programme able to identify whether users of an adult website are telling the truth about their age and gender. They found that nearly 40 per cent of users could be lying about their age, and 25 per cent about their gender.

A fake online profile – often referred to as a ‘sock puppet’ – may be set up to protect anonymity, give the appearance of popularity to the user’s main account, or engage in fake online romances.

Social media users who make up information about themselves – often inventing a whole new persona – are known as catfish.

In order to better identify catfish lurking among the genuine users on social networks, a team of computer scientists based at the University of Edinburgh trained a neural network to spot accounts being held by users lying about their age or gender.

The researchers chose to focus their research on the adult content website, PornHub, which as well as hosting explicit videos has a number of social features that allow users to create profiles and engage with each other and with content. These ‘Porn 2.0’ website are targeted frequently by catfish using fake identities in order to connect with other users.

The models were provided with data collected from 5000 verified public profiles. This was used to train the programme to estimate the age and gender of a user with high accuracy, picking up on clues in their writing style and network activity.

It was then used to estimate the age and gender of unverified account holders. The Edinburgh researchers found that catfishing on adult websites is very widespread: 38 per cent of PornHub users may be lying about their age, and 25 per cent about their gender, with women being more likely to lie than men.

While men pretending to be women consistently lie about being younger, women pretending to be men select from a wider range of ages.

“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,” said Dr Walid Magdy, based at the University of Edinburgh’s School of Informatics.

“We hope that our development will lead to useful tools to flag dishonest users and keep social networks of all kinds safe.”

The technology could be helpful in keeping users of adult websites and other social networks safe, for instance by identifying and removing underage users, the authors suggest.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them