Innovative project will use AI to help policymakers to recognise and counter disinformation
The ‘Polinoculate’ project will develop a freely accessible AI ‘coach’ to give policy stakeholders the tools to counter the rising problem of information manipulation, whilst safeguarding freedom of expression.
Whilst the EU has advanced digital regulation, its effectiveness is unclear in the face of rising disinformation. This includes, for example, false claims about climate change and migrants, as well as reporting politicians' statements out of context or with unfair editing, and distraction information that seeks to overwhelm a message with irrelevant facts or spurious counter-claims.
The AI ‘coach’ that the international team of researchers on the Polinoculate project will develop will employ ‘psychological inoculation techniques’ to equip policymakers to recognise, resist and counter manipulated information.
The coaching will focus especially on targeting disinformation that has been generated by AI, which is the greatest challenge to identify, as it can vary the format of messages almost infinitely to avoid detection.
The team at the University of Sheffield is made up of Professor Mike Thelwall and Dr Irini Katsirea from the School of Information, Journalism and Communication, and Dr Carolina Scarton and Dr Xingyi Song from the School of Computer Science.
Professor Thelwall will identify and map the scale and content of current disinformation across social media, whilst Dr Katsirea will analyse the legal aspects of misinformation in Europe. The team led by Dr Scarton will build AI-driven agents and a web dashboard for monitoring and analysing misinformation in real time.
Other teams from the University of Turku (Finland), University of Bristol, Universtitat Passau (Germany) and Eotvos Lorand Tudomanyegyetem (Budapest, Hungary) will cover other areas of the project including building a training syllabus for policymakers, understanding the political opposition countering misinformation, and developing detection recommendations. Global thinktank Globsec will support the dissemination of the project’s results and engage policymakers in Polinoculate's co-creation activities.
Mike Thelwall, Professor of Data Science, said: "I am very happy to get started on developing methods to protect policymakers against misinformation and disinformation. We have a very multidisciplinary team of information scientists, computer scientists, psychologists, sociologists and political scientists that is well placed to understand and respond to the new challenges from the AI era."
Dr Scarton, Senior Lecturer in Natural Language Processing, added: “Information manipulation is one of the biggest risks to democracy, whilst some ways of preventing it may result in reduced freedom of expression. Polinoculate addresses this challenge by bringing together renowned multidisciplinary experts and stakeholders to co-create tools and techniques for disinformation analysis. I am looking forward to working on this project and collaborating with our consortium.”
The project will run until August 2029.
Find out more on the project website here.