Dr. Dmitri Bershadskyy
Dr. Dmitri Bershadskyy
2025
Peer-reviewed journal article
A laboratory experiment on using different financial-incentivization schemes in software-engineering experimentation
Bershadskyy, Dmitri; Krüger, Jacob; Çalıklı, Gül; Otto, Siegmar; Zabel-Öztürk, Sarah; Greif, Jannik; Heyer, Robert
In: PeerJ computer science - London : PeerJ, Ltd., Bd. 11 (2025), Artikel e2650 , insges. 34 S.
Article in conference proceedings
Comparing OpenFace and deep learning models for deception detection in video calls
Dinges, Laslo; Fiedler, Marc-André; Al-Hamadi, Ayoub; Bershadskyy, Dmitri; Weimann, Joachim
In: 2025 14th International Symposium on Image and Signal Processing and Analysis (ISPA) - [Piscataway, NJ] : IEEE, S. 231-236 [Konferenz: 2025 14th International Symposium on Image and Signal Processing and Analysis, ISPA, Coimbra, Portugal, 29-31 October 2025]
Non-peer-reviewed journal article
Lie Against AI - revealing private information through AI in an economic experiment
Bershadskyy, Dmitri; Dinges, Laslo; Fiedler, Marc-André; Greif, Jannik; Al-Hamadi, Ayoub; Ostermaier, Nina; Weimann, Joachim
In: SSRN eLibrary - [Erscheinungsort nicht ermittelbar] : Social Science Electronic Publ. . - 2025, insges. 29 S.
2024
Peer-reviewed journal article
Choosing a victim you know - introducing communication to the mobbing game
Bershadskyy, Dmitri; Seidel, Alexandra
In: Journal of behavioral and experimental economics - Amsterdam [u.a.] : Elsevier, Bd. 112 (2024), S. 1-7, Artikel 102265
Exploring facial cues - automated deception detection using artificial intelligence
Dinges, Laslo; Fiedler, Marc-André; Hamadi, al- Ayoub; Hempel, Thorsten; Abdelrahman, Ahmed; Weimann, Joachim; Bershadskyy, Dmitri; Steiner, Johann
In: Neural computing & applications - London : Springer, Bd. 36 (2024), Heft 24, S. 14857-14883
MTVE - Magdeburg tool for video experiments
Bershadskyy, Dmitri; Ghadwal, Sunil; Greif, Jannik
In: Journal of the Economic Science Association - Cambridge : Cambridge University Press, Bd. 10 (2024), Heft 2, S. 609-619
Guidelines for using financial incentives in software-engineering experimentation
Krüger, Jacob; Çalıklı, Gül; Bershadskyy, Dmitri; Otto, Siegmar; Zabel, Sarah; Heyer, Robert
In: Empirical software engineering - Dordrecht [u.a.] : Springer Science + Business Media B.V, Bd. 29 (2024), Heft 5, Artikel 135, insges. 53 S.
Experimental economics for machine learning - a methodological contribution on lie detection
Bershadskyy, Dmitri; Dinges, Laslo; Fiedler, Marc-André; Hamadi, al- Ayoub; Ostermaier, Nina; Weimann, Joachim
In: PLOS ONE - San Francisco, California, US : PLOS, Bd. 19 (2024), Heft 12, Artikel e0314806, insges. 19 S.
Non-peer-reviewed journal article
ChatGPT's financial discrimination between rich and poor - misaligned with human behavior and expectations
Bershadskyy, Dmitri; Sachs, Florian E.; Weimann, Joachim
In: Arxiv - Ithaca, NY : Cornell University . - 2024
2023
Book chapter
Uncovering lies - deception detection in a rolling-dice experiment
Dinges, Laslo; Fiedler, Marc-André; Al-Hamadi, Ayoub; Abdelrahman, Ahmed A.; Weimann, Joachim; Bershadskyy, Dmitri
In: Image Analysis and Processing – ICIAP 2023 , 1st ed. 2023. - Cham : Springer Nature Switzerland ; Foresti, Gian Luca, S. 293-303 - ( Lecture notes in computer science; volume 14233)
Peer-reviewed journal article
Reverberation effect of communication in a public goods game
Bershadskyy, Dmitri
In: PLOS ONE - San Francisco, California, US : PLOS, Bd. 18 (2023), Heft 2, Artikel e0281633, insges. 20 S.
Collective bargaining in a shrinking group game - the role of information and communication
Bershadskyy, Dmitri; Sachs, Florian E.; Weimann, Joachim
In: Journal of economic behavior & organization - Amsterdam [u.a.] : Elsevier, Bd. 209 (2023), S. 391-410
Non-peer-reviewed journal article
Automated deception detection from videos - using end-to-end learning based high-level features and classivication approaches
Dinges, Laslo; Al-Hamadi, Ayoub; Hempel, Thorsten; Abdelrahman, Ahmed; Weimann, Joachim; Bershadskyy, Dmitri
In: De.arxiv.org - [Erscheinungsort nicht ermittelbar] : Arxiv.org . - 2023, Artikel 2307.06625, insges. 29 S.
Experimental economics for machine learning - a methodological contribution
Bershadskyy, Dmitri; Dinges, Laslo; Fiedler, Marc-André; Hamadi, Ayoub; Ostermaier, Nina; Weimann, Joachim
In: Magdeburg: Otto-von-Guericke-Universität Magdeburg: Fakultät für Wirtschaftswissenschaft, 2023, 1 Online-Ressource (27 Seiten, 0,68 MB) - (Working paper series; Otto von Guericke Universität Magdeburg, Faculty of Economics and Management; 2023, no. 08)
Registered report: A laboratory experiment on using different financial-incentivization schemes in software-engineering experimentation
Krüger, Jacob; Çalıklı, Gül; Bershadskyy, Dmitri; Heyer, Robert; Zabel, Sarah; Otto, Siegmar
In: De.arxiv.org - [S.l.] : Arxiv.org . - 2023, Artikel 2202.10985, insges. 10 S.
2022
Non-peer-reviewed journal article
MTV - Magdeburg Tool for Videoconferences
Bershadskyy, Dmitri; Ghadwal, Sunil; Greif, Jannik
In: Magdeburg: Otto-von-Guericke-Universität Magdeburg: Fakultät für Wirtschaftswissenschaft, 2022, 1 Online-Ressource (10 Seiten, 0,71 MB) - (Working paper series; Otto von Guericke Universität Magdeburg, Faculty of Economics and Management; 2022, no. 9)
2019
Peer-reviewed journal article
Predicting group contribution behaviour in a public goods game from face-to-face communication
Othman, Ehsan; Saxen, Frerk; Bershadskyy, Dmitri; Werner, Philipp; Al-Hamadi, Ayoub; Weimann, Joachim
In: Sensors - Basel : MDPI - Volume 19 (2019), 12, Artikelnummer 2786 [Special Issue: Sensors for affective computing and sentiment analysis]
Current projects
Acceptance of Algorithmic Advice
Duration: 01.11.2023 to 31.03.2026
Advancing digitalization makes it possible to increasingly replace or support human decisions with powerful algorithms. Nevertheless, people often avoid using such algorithms, a phenomenon called algorithm aversion. While previous studies have investigated various influencing factors, it remains unclear how the initial human-computer interaction influences this aversion. Since attitudes towards algorithms can emerge and change during the interaction, targeting and analyzing these interactions offers an opportunity to develop strategies to reduce aversion. This will be investigated experimentally by analyzing how the way information about the algorithm is presented influences perception.
This text was translated with DeepL on 26/02/2026
Completed projects
Behavioral preferences against AI
Duration: 01.12.2023 to 31.12.2025
ChatGPT has simplified the application of machine learning and is used in a variety of ways, for example for consulting, coding or summarizing information. However, its potential extends to negotiation situations. To investigate this, we use a laboratory experiment with the ultimatum game, in which a provider makes a monetary offer to a recipient. In our design, ChatGPT assumes the role of the offerer, and we vary the wealth of the recipients.
This text was translated with DeepL on 26/02/2026
Economics and Rasch Model from Psychology
Duration: 01.06.2024 to 31.12.2025
The way in which people make decisions depends on numerous factors. In this project, we combine the theoretical and experimental foundations from economics (Expected Utility Theory, Behavioral Economics) with the application of the Rasch model from psychology.
This text was translated with DeepL on 26/02/2026
"Financial Incentives in Software Engineering"
Duration: 01.12.2022 to 30.06.2025
Empirical studies with human participants (e.g. controlled experiments) are well-established methods in software engineering (SE) research to understand the activities of developers or the advantages and disadvantages of a technique, tool or practice. There are various guidelines and recommendations for designing and conducting different types of empirical studies in SE. However, the use of financial incentives (i.e. paying participants to compensate their effort and improve the validity of a study) is rarely mentioned.
In this project, we analyze and discuss the use of financial incentives for SE experiments in order to derive appropriate guidelines and recommendations for researchers. In particular, we propose how to extend the current state of the art and create a better understanding of when and how incentives should be used.
This text was translated with DeepL on 26/02/2026
The impact of using AI-powered technology to detect lies in negotiations
Duration: 01.01.2022 to 31.12.2024
The increasing digitalization of social and economic interactions is proceeding at a considerable speed. Research on digitalization processes should reconcile two areas of knowledge that are usually examined separately: Firstly, the question of technical development and secondly, the question of the effects of this development on human behavior. In the project applied for here, an attempt will be made to combine both perspectives in an interdisciplinary approach, whereby the focus is on behavioral analysis, but the technical components are still strongly represented. The use case chosen for this type of analysis of digitization processes is the phenomenon of asymmetric information. Specifically, we are investigating the extent to which the paradigm of asymmetric information distribution has become at least partially obsolete through the use of AI technologies. In our interdisciplinary project, instead of waiting for the technological development in the field of machine lie detection, we would like to contribute to technological progress and at the same time experimentally investigate the possible social consequences of this technology: Economics (WW) and Neuro-Information Technology (NIT). The identification of private information plays a major role in both areas, but is viewed from different perspectives. While economic analysis focuses on the role and importance of private information in negotiation situations, NIT focuses on the feasibility and quality of automated recognition of personal characteristics, with the main aim of answering two questions. (1) Does the existence of such AI technologies lead to market selection effects? (2) Under which circumstances will actors prefer a market design that uses or omits AI-based technology? As key sub-objectives, we want to find out how well lying can be detected by machines in our negotiation context and how the subjects deal with the information provided by this AI technology, thus addressing the economic consequences of digitization in the field of market design. The aim is to investigate how the players select themselves into corresponding submarkets and whether the markets come into existence at all. This research question is of high practical and scientific relevance. The use of AI will increase significantly over the next few years and will appear in more and more applications. In view of this development, it is of great social and scientific interest to know how the use of such technologies influences the behavior of the actors concerned.
This text was translated with DeepL on 26/02/2026