Although disinformation has been a recognized problem for some time now and is actively dealt with at national and international level, it continues to gain in explosiveness on a daily basis. From disinformation-led voter manipulation, to willful influence on pandemic response, to targeted war propaganda. The latter has also drawn attention to a new phenomenon: hybrid threats. This abstract aggregation of different threats thwarts traditional responsibilities of Austrian security agencies, in that by definition they are carried out by state or non-state actors, through a combination of overt and covert military and non-military means, and fall within the purview of either the BMI (e.g. cyber-security, cyber-crime), BKA (e.g. disinformation, public security), BMLV (e.g. cyber-defense). A concerted attack using Hybrid Threats therefore requires the cooperation of security agencies, or the facilitation of cooperation or the exchange of analysis results and counter strategies. This is the starting point of HYBRIS.
HYBRIS addresses hybrid threats that have the intent to influence beliefs and attitudes of selected audiences through online coordinated operations, mobilize them to act, and subsequently compromise physical and digital infrastructure. Disinformation is an essential component of hybrid threats. Disinformation campaigns are often aimed at stoking fears in the population, which can result in various consequences. Whether or not messages are misinformation, it is important for security agencies to identify whether reactions to messages on social media and other news channels, or potentially subsequently organized actions, pose a threat to people or critical infrastructure. Only when it comes to taking appropriate action, in the case of the spread of disinformation, is it also a matter of counteracting a threatening development by creating and disseminating counter-statements based on facts and trustworthy information.
It is particularly important to respond quickly to this type of news trend. Therefore, preparation of the content of possible counterstatements must begin while they are still in the making. One of the main challenges here is to ensure a comprehensive overview of the currently prevailing trends. The difficulty here is that information is exchanged at high speed via an unmanageable number of information channels, and it is impossible to obtain an overview by manual sifting. In addition, the trustworthiness of the information must be assessed in the course of matching different information sources and modalities. For these tasks, there is an urgent need to support security agencies with automated and AI-based systems for information sifting and overview generation.
The challenge for security agencies in this context is to cope with the information overload, which is accompanied by an ever-increasing data load.
A primary goal of HYBRIS is to structure unstructured data with the help of artificial intelligence in such a way that the broadest possible overview of essential information can be provided. For this purpose, approaches for the detection of relevant narratives such as "suspected fuel shortage", which poses a threat to the critical infrastructure "traffic" due to the impact "congestion in front of gas stations", will be explored. By recognizing narratives and their context, security forces gain the latitude to respond accordingly. Content structured according to topics, narratives or categories is to be automatically evaluated, e.g. with regard to disinformation, so that a classification with regard to hybrid threats is possible, which additionally extends the decision-making scope of the authorities.
HYBRIS is to be a prototype for a national data intelligence platform, which will be used to create a decision-making basis for improving the effectiveness and resilience of the Austrian security authorities.
Contact
Dr. Alexander Schindler
AIT Austrian Institute of Technology GmbH
Giefinggasse 4,
1210 Wien, Österreich
+43 50550 2902
Alexander.Schindler@ait.ac.at
Homepage: https://www.ait.ac.at
Oder: https://www.ait.ac.at/themen/data-science-artificial-intelligence
Project parnters
Research Institute AG & Co KG Amundsenstraße 9, 1170 Wien
Kontakt: Christof Tschohl christof.tschohl@researchinstitute.at
Thinkers.ai
Maria-Jacobi-Gasse 1, 1030 Wien, Österreich
Kontakt: Isabell Claus isabell.claus@thinkers.ai
TU-Wien
Karlsplatz 13, 1040 Wien, Österreich
Kontakt: Siegfried Höfinger siegfried.hoefinger@tuwien.ac.at
Universität für Bodenkultur Wien
Peter-Jordan-Straße 62, 1190 Wien, Österreich
Kontakt: Markus Stöhr markus.stoehr@tuwien.ac.at
Artificial Researcher IT GmbH
Taubstummengasse 11, 1040 Wien, Österreich
Kontakt: Linda Andersson linda.andersson@artificialresearcher.com
Bundesministerium für Landesverteidigung
Roßauer Lände 1, 1090 Wien, Österreich
Kontakt: Michael Suker michael.suker@bmlv.gv.at