A Comparison of Fake News Detecting and Fact-Checking AI Based Solutions
Andrej Školkay
Juraj Filin
ABSTRACT
Scientifi c objective of this paper is to analyse how advanced are Artifi cial Intelligence (AI) tools to fight successfully information disorder. More specifi cally, this is an overview and ranking on existing tools based on AI in this specifi c area. Research method is comparative analytics. We compare the most developed and publicly available fake-news detecting and fact-checking AI based solutions (intelligent machines). The comparison is based on two key parameters: accuracy and comprehensiveness. Results and conclusions: Analyse show that a third of the examined AI systems are, in terms of comprehensiveness, in the top category, while the majority are in the medium category. As far as accuracy is concerned, very few AI machine developers are interested in providing further details about their products and functionalities for studies such as ours which raises suspicions about their actual performance. Surprisingly, one of the most discussed AI systems among EU leaders seems to actually belong to the least developed. Cognitive value: There is a need for a larger and more detailed study with involvement of AI specialists who would be able, and allowed, to test all available AI machines with their key features and functionalities.
This is partial deliverable of the H2020 CSA Project: COMPACT: From Research To Policy Through Raising Awareness of the State of the Art on Social Media and Convergence, Project Number 762128.
see more http://compact-media.eu/
