KInIT created a tool that people can use to flag misinformation
Bratislava, December 17, 2021 – Since the fight against misinformation is a superhuman task, Kempelen’s Institute of Intelligent Technologies (KInIT) wants to use artificial intelligence in it. However, artificial intelligence algorithms first need to learn to distinguish misinformation from humans. KInIT has, therefore, created a tool which people can use to flag (annotate) manipulative and fake articles and posts on social networks. Just download the extension to your browser to annotate.
KInIT Označovač (translators note: a marker) will serve as the first step to artificial intelligence learning about Slovak misinformation. Such artificial intelligence is used abroad, for example, as help and support for fact-checkers. At present, however, there is very little data available on Slovak fake news websites. Therefore, KInIT invites all those who want to contribute to the collection of the necessary data to try it out at www.oznacuj-dezinfo.kinit.sk.
“Since this is the first such Slovak initiative, every annotation sent is good for us. The more annotations we get from people, the better the results of the artificial intelligence for which these annotations will be used, ”says KInIT researcher Branislav Pecher.
Learning the algorithms of artificial intelligence itself works on the basis of patterns that reveal to them what something looks like. Article annotations will serve as signals indicating that we may be dealing with misinformation. These will then be used as patterns for artificial intelligence to learn to detect such signals automatically.
The tagging process itself is easy to use. Annotators will find a tool on the website www.oznacuj-dezinfo.kinit.sk, as well as all the necessary instructions and a tutorial for marking problematic content.
The tool is available to anyone, but it can be most useful for those who are bothered by fake and manipulated news. These can be people with a higher level of information literacy, disinformation experts, journalists, but also experts on various topics on which disinformation spreads (from astronomers to health professionals).
The project was supported by the O2 Fair Foundation as part of the call “Vráťme spolu rozum na internet” (Let’s restore common sense on the Internet).
In recent days, KInIT has been testing the tool and collecting feedback. Annotators who have been involved so far have appreciated such activity and believe that they can do something about a huge amount of manipulative content. The tool is now publicly available and KInIT researchers are encouraging people to use it.
“There is no lower or upper limit on the number of annotations or time. We will be happy if as many people as possible get involved. The more data, the better, ”adds Branislav Pecher.
KInIT has extensive experience with misinformation. It is part of the CEDMO hub (Central European Digital Media Observatory), which associates 8 professional partners from the Czech Republic, Poland and Slovakia. Within it, KInIT is engaged in research and development of tools based on artificial intelligence, which support fact checking.
In addition, it develops Monant and Fireant platforms for monitoring, analyzing and annotating misinformation through artificial intelligence models. The research institute has also published several scholarly articles on misinformation. One of them, which focused on the audit of filter bubbles on YouTube, was awarded as the best article at the prestigious international conference ACM Recommender Systems Conference (RecSys) 2021.