The subject of the project is to create a system for verifying information sources and detecting fake news involving the use of computer-assisted argumentation and a reusable knowledge base to protect the public from manipulation. The system will collect information by using users of a browser plug-in to extract information for checking, a browser application for discussion and browsing the knowledge base, a widget and a chatbot for discussion and automation of informing users about detected fake news and popular misinformation in public discourse.

The process of fact-checking relies on algorithmic techniques (Deep Learning, NLP, web crawling, semantic networks) and collective intelligence, users of the system do not need to be fact-checking experts. The system will allow the integration of scattered discussion fragments into a coherent database of validated and reusable arguments, which will be used to evaluate information. Algorithms will be created to classify argumentation patterns, a resolving algorithm that ensures that conflicting information is removed as objectively as possible, and argument mining algorithms that will help automate distributed argumentation processes. In line with an ethical approach to AI development, the system supports human capabilities, transparently and explainably presents justifications for accepted classifications using argumentation graphs, and enables simple error correction using natural language.

The SaaS freemium implementation will provide free access for all users using the public knowledge base (open access). There will be a fee for the use of private discussions. The system will be integrated into Swarmcheck's existing activities - consulting services in group decision support, strategy development, qualitative research and participatory processes, as well as popular-science and educational activities with partner communities.