SciNet – Network Analysis Tool to automatically rate web domains on their scientific credibility

Grantee: Science Feedback / Grant: €79,645
  • Priority Area Boosting Fact-checking Activities in Europe
  • Funding Round / Action Types 2nd Funding Round – Scale-up Projects
  • Year 2022, 2023
  • Country France
  • Project Status Ongoing

The vastness of the online information space and its fragmentation across countries and languages make a manual evaluation of each domain’s risk of carrying misinformation impossible. While high-profile outlets which repeatedly publish or amplify misinformation might be well-identified by teams specialized in source evaluation, countless smaller ones fly below the radar. This is especially true for websites publishing in lesser-spoken languages across the EU.Yet, these websites matter: misinformation researchers might want to track what messages they are carrying, social media platforms might want to inform their users of the risk of misinformation before they share a link, brands might not want their ads to appear on those websites.The sheer volume of the task of evaluating tens or hundreds of thousands of websites across the EU’s 27 countries and 24 official languages means it cannot be done by humans alone.The SCENT (Scientific Credibility Evaluation NeTwork) project aims to fill this gap. First, our trained scientists will manually review websites and evaluate the risk that they carry misinformation. This data will then be used to train AI algorithms capable of automating this evaluation, so as to provide the volume needed to make a difference at Internet-scale.The end result will be a list of tens or hundreds of thousands of websites across all EU languages, with a misinformation risk assessment estimate attached to each. A portion of them will have been rated by our human experts, but the majority will have been rated by our AI tool.

Updated on 03 october 2022

Cookies settings

Cookies Selection

The Calouste Gulbenkian Foundation uses cookies to improve your browsing experience, security, and its website performance. The Foundation may also use cookies to share information on social media and to display messages and advertisements personalised to your interests, both on our website and in others.