about us research technology projects education collaborations news scopus intranet
about us research technology projects education collaborations news scopus intranet


condividi: facebook_icon twitter_icon

Be careful what you read on social networks: between fake profiles and fake content, getting (well) informed is not easy

Marinella Petrocchi studies the spread of fake news and the role of social bots in the network ecosystem, so as to create useful tools for users

We hear it said over and over again: look out, not everything circulating on the Internet is true. We face it when election time comes round, when we try to get an idea of who to vote for, when we read news that seems unbelievable or simply when we tend to place too much trust in a post read on Facebook. The phenomenon is so widespread that entire lines of research have been created in the field of the Internet dedicated to studying and identifying all that is false on the net. Marinella Petrocchi, Senior Researcher at the Trust, Security and Privacy research unit of the IIT-CNR, works in that field.

Persons or robots?

“I have studied social bots for a long time, in collaboration with Maurizio Tesconi and Stefano Cresci” (also researchers at IIT-CNR, recently seen on our live “Il mondo dei bot”)”. Social bots are algorithms programmed to perform a certain task and hide behind a person’s social profile. They are to all intents and purposes robots, even if they don’t have circuits and gears, and they can blend in very well on social networks like Twitter. “They are often programmed for harmless and indeed useful purposes, like the automatic dissemination of news, but there is also a real market for fake accounts, with credible profiles and posing as real people”, she explains. Whoever wants to expand their following, from celebrities to politicians, and including the big brands, is interested in buying these profiles.

Petrocchi and colleagues have worked on detection, developing techniques to automatically identify fake profiles. They were among the first in the world to deal with the subject, starting from the intuition of studying profiles in groups rather than individually. “Bots often act in a coordinated manner often act in a coordinated manner, spreading misleading messages” (the bot squads phenomenon, wich Petrocchi and her colleagues have studied with the 2020 paper “The role of bot squads in the political propaganda on Twitter). “It has been well proven that they have played a role in the main electoral campaigns, including those of Trump and Macron”.

Social manipulation

On the net, it’s not just the people that are fake, but the news too. “For a couple of years now I have been dealing with disinformation and misinformation, terms which, with different nuances, describe the proliferation of false content on social media”. The first term is used to indicate an untruthful content disseminated with the intention of contaminating the public debate, while the second does not presuppose malicious purposes, only the lack of awareness of whoever is spreading the content.

Petrocchi has an ongoing project in collaboration with the IMT School for Advanced Studies Lucca, Toffee, in which computer scientists, statistical physicists and mathematicians join forces to monitor the activity of bots on the Internet. In particular, the researchers look for anomalies in the diffusion of news. “These are strategies put in place by real people, who use groups of bots to amplify the diffusion of scarcely reputable news. For example, in a dataset of tweets on migrants in Italian, it emerged that some communities of users, mostly belonging to the conservative political area, managed to get their arguments across thanks to teams of bots”.

The final goal of the project is the creation of a platform that can guide those who obtain information on the net, reporting the presence of bots and indicating the most reliable sources on the social networks.

Virtual prisons

Finally, a topic very dear to Petrocchi concerns the opportunities offered by social networks. “I have often wondered if the Internet is a resource or a prison”, she says. I’m thinking in particular of the high risk of those who often use social networks to get information and make new friends, finding themselves inside a bubble. It’s something we experience everyday as common users. Algorithms tend to show results that are related to your interests, what you like, and what you would like to read. In recent years we have studied the phenomenon and how to counter it in the Mib project ”.

Marinella Petrocchi was interviewed a few weeks ago by the RaiPlay program ConverseRai.

share: facebook_icon twitter_icon
all news