AI in the service of truth

One of the biggest problems democracy is facing today is disinformation spread and the challenge of acquiring the right information and facts. The rise of social media has transformed the way we consume information and facts and at the same time has opened up for an extremely cheap and efficient way of disinformation spread in a viral manner. Beyond social media, media in general could be polarized or biased. It is frequent to run into statements and claims in the news that are incorrect, twisted, or lack evidence. Also, some times one reads pure citations from politicians for instance, and it is of highest interest to verify their claims and statements and relate those to previous statements to assess the truthfulness and consistency. Furthermore, videos have been a strong evidence to capture facts and statements. However, this might not be reliable in 5 years from now as fake videos produced with machine learning are getting strikingly near to realistic.

The aim of the project is to build a toolkit that can be used for detecting disinformation, for verifying the correctness of claims, and for quantifying the quality of references as an attempt to counteract fake news and spread of disinformation. This includes functionalities such as detecting disinformation, propaganda, and polarized content, as well as verification of the correctness of claims and statements in texts as well as in videos. The software could be used by journalists, policy makers, and society in general. Such a toolbox will empower people in general with tools that help them to check their facts. One positive consequence will be that journalists, politicians, and people in general will be more careful and refrain from making statements that lack support.