Explainable AI

The goal of this research effort is to develop novel approaches and tools to detect various types of intent in social media campaigns.
Meanwhile it is considered credible that during the 2016 US presidential election foreign powers have started impactful media campaigns aiming at influencing the opinion of voters in ways which are considered non-legitimate, see e.g. https://www.washingtonpost.com/politics/through-email-leaks-and-propaganda-russians-sought-to-elect-trump-mueller-finds/2019/04/18/109ddf74-571b-11e9-814f-e2f46684196e_story.html https://www.justice.gov/storage/report.pdf .
Similar campaigns are thought to be run on smaller scale for commercial products, e.g. cosmetics, medical supplements https://www.hsa.gov.sg/announcements/press-release/healthproductsmakingfraudulentcovid19claims , or against persons of public interest which may be targeted with negative campaigning without political motives, e.g. Venus Williams in the aftermath of https://www.straitstimes.com/sport/tennis/tennis-venus-williams-settles-lawsuit-over-fatal-florida-car-crash .

Here we look into ways to discover campaigns as they are starting, their likely topics, their sentiment against persons or towards products of interest, as well as their expected development over time.