Hot off the press

Machine learning (ML) applications face many new, hardly predictable aspects in their production environments. Detecting new aspects in an ML production environment and understanding their impacts on the ML application is crucial if organizations are to ensure ML applications’ functionality. A monitoring entity is essential if one is to monitor ML applications in their production environments, to both continually minimize risks and improve ML application’s performance. But existing monitoring approaches are struggling to deal with specifics that arise from ML applications.

In our recently published research, we aimed at deriving monitoring practices and providing a holistic view over the required steps in successful ML applications monitoring. Since there has been little research on this topic, we followed a qualitative research approach, i.e., we conducted an interview study combined with a multivocal literature review. Thus, we provide a theoretical framework of an ML-enabled agent in its production environment, five characteristics of ML applications’ production environments and 17 monitoring practices – 14 practices arranged sequentially on a typical quality management cycle and three cross-sectional practices. To outline the ML specifics that arise in monitoring ML applications, we investigate the five ML production environment characteristics’ influences on the ML monitoring practices.

I am happy that our paper “What Gets Measured Gets Improved: Monitoring Machine Learning Applications in their Production Environments” has been accepted for publication in IEEE Access and is now available online (Open Access): https://ieeexplore.ieee.org/document/10886935

Corporate bonds are an attractive option for corporate financing. However, current bond markets face many challenges and inefficiencies, resulting in high transaction costs (TAC). In recent years, technological advancements like blockchain technology have enabled the possibility of reducing TAC in bond markets. Even though practice experiments with such solutions, academic literature lacks generic design knowledge under the TAC lens to design blockchain-based bonds.

Thus, in a recent research project, we followed the design science research (DSR) paradigm to design and develop a bond prototype using the Ethereum blockchain protocol. Our results highlight the capability of blockchain-based bond markets to reduce TAC in the three dimensions of asset specificity, uncertainty, and transaction frequency. Further, our research provides design principles to contribute to both practice and the academic discourse on developing blockchain-based bond markets with reduced TAC.

I am happy that our paper “Designing the future of bond markets: Reducing transaction costs through tokenization" has been accepted for publication in Electronic Markets and is now available online (Open Access): https://link.springer.com/article/10.1007/s12525-025-00753-3.

I am happy to announce the release of our new book, »Decentralization Technologies: Financial Sector in Change«, edited together with my esteemed colleagues Gilbert Fridgen, Dr. Tobias Guggenberger, and Johannes Sedlmeir. This volume is now part of the Financial Innovation and Technology (FIT) series by Springer Nature. In this edited collection, we look at the transformative role of decentralization technologies in reshaping the financial services industry. We bridge the gap between technological innovation, regulatory requirements, and the evolving expectations of consumers in increasingly competitive and data-driven financial markets.

I am incredibly proud of the insights shared by our contributors and the practical applications this book outlines. Our book is aimed at researchers, practitioners and anyone interested in blockchain, financial markets, data governance and digital transformation. For more details or to get your copy, visit the publisher's website. Many universities have a SpringerLink subscription with which you can access the book free of charge from your university network. A huge thank you to all contributors and supporters who made this project possible. Let’s continue to drive innovation and shape the future of finance together!

AI strategy and human computer interaction (HCI) are mostly independently considered fields in both research and practice, limiting sustainable and scaled value creation. In our Electronic Markets special section “AI-enabled information systems: Teaming up with intelligent agents in networked business”, we shed more light on the permeation of strategic considerations, such as an organization’s AI ambition, to the individual employees and vice versa.

In our editorial paper, we introduce the ecological work systems framework and take a closer look at three permeation issues:

1. Missing architecture perspective
2. Missing operative glue
3. Static transformation steering methods

This special issue is a testament to the value of the exchange between research and practice. By collecting relevant issues from the field, we were able to outline future research paths that actually solve problems for business and society.

Find out in our new white paper.

The introduction of the European Digital Identity Wallet (EUDI-Wallet) is the basis for the creation of a digital identity ecosystem in the European Union. This development is closely linked to the amendment of the Electronic Identification, Authentication and Trust Services (eIDAS) Regulation which facilitates the cross-border recognition of recognition of digital identities and trust services within the EU. The Fraunhofer Institute for Applied Information Technology (FIT) offers with this white paper a comprehensive analysis of the use cases, potential uses and challenges of the challenges of the EUDI wallet for companies.

The eIDAS Regulation aims to create a Europe-wide basis for digital identities, which are particularly essential when using cross-border electronic services. The EUDI wallet is intended to serve as a central digital wallet in which verifiable data is stored securely. The EU member states are obliged to offer a free EUDI wallet solution by 2026. This should ensure broad dissemination of the ecosystem by 2030. To this end, the technical foundations and practical applicability of the wallet are currently being tested in four European pilot projects.

The white paper highlights various areas of application for the EUDI wallet. In the financial sector, the EUDI wallet can facilitate compliance with legal requirements by enabling automated and secure verification of digital identities. This can save effort and costs for the financial institutions using it. In the retail sector, the EUDI Wallet can be used for the purchase of goods that are subject to certain conditions, such as age verification or personalized digital goods. This can improve the efficiency and security of transactions. There are also promising potential applications in the healthcare sector. The integration of e-prescriptions into the wallet can simplify the purchase of prescription drugs. In addition, medical product passports and electronic patient records can be managed efficiently. For public administration, the wallet represents an opportunity to consistently implement the digitalization of processes, which can relieve the burden on companies and individuals in particular who frequently interact with authorities.

Please read here more about the European Digital Identity Wallet in our recently published white paper (in German).

In the aftermath of the media attention given to the Cum/Ex structures the German government in their 2021 coalition agreement decided to further pursue the prevention of dividend arbitrage transactions, i.e. abusive arrangements in the context of dividend taxation, and to explore the possible use of modern technologies to achieve this objective. Dividend arbitrage transactions are based on the special taxation procedure for capital gains tax. Capital gains tax is initially withheld in a standardised manner at the level of the withholding entity and then, if required, corrected at the level of the taxpayer. This taxation procedure in combination with the complex depository system of the financial market is abused so as to impede the provision of the information required for lawful taxation.

Blockchain technology, which has already found its way into various financial applications, was explicitly mentioned in the coalition agreement as a technology for preventing dividend arbitrage. A blockchain is a decentralised register of chronologically ordered transactions aggregated in blocks that can ensure the availability and immutability of data and the integrity of the system independently of a central party. This study analyses the extent to which these properties can be used to prevent tax avoidance schemes. Furthermore, it takes into account legal, economic, and technical requirements as well as system costs to propose a practically feasible solution.

The primary result of the study is a concept that attempts to address both basic types of tax avoidance schemes: On the one hand, tax schemes with the aim of multiple tax refunds (e.g. cum/ex) are prevented by a token-based solution in which tokens on the blockchain issued by the tax administration and to be redeemed upon tax refund ensure that the amount of refunded tax cannot exceed the amount that was originally paid. On the other hand, additional information on the tokens makes it easier to uncover and trace tax avoidance schemes aimed at misusing tax refund claims or avoiding definitive tax charges (e.g. cum/cum).

By presenting an explicit solution concept the study demonstrates that blockchain technology can already be used to prevent excessive tax refunds. By providing traceable and tamper-proof information the technology also has the potential to create benefits in the fight against other tax avoidance schemes. However, maximising these benefits would require using blockchain technology as base for the entire financial infrastructure, especially for securities book entry, trading, and settlement. While this is considered unrealistic in the short term, current regulatory, technological and social advances may set a path to enable this in the future.

As a result, the study not only shows the current possibilities of blockchain technology to prevent tax avoidance in the form of dividend arbitrage, but also provides an outlook for the future, offering an additional incentive to continue and even intensify efforts to use blockchain technology for the financial infrastructure.

Please access the complete study report here (in German).

The increasing capabilities of Artificial Intelligence (AI) raise concerns about the risks associated with the technology. The European Union, therefore, proposed the Artificial Intelligence Act aiming to mitigate the risks of AI by fostering their safety and transparency. However, there is controversial debate about its impact on AI innovation. While the AI Act aims to provide legal certainty guiding innovation, the criticism refers to exaggerated bureaucratic burden such as transparency requirements impeding innovation.

Based on a multivocal literature review, we examined the impact of the AI Act's transparency requirements on patenting as a means for AI innovation. Our results indicate that the transparency requirements do not necessarily hinder the patentability of AI innovations. Instead, existing concerns primarily rely on uncertainties within key terms of the AI Act. Accordingly, we propose an improvement suggestion focusing on resolving existing uncertainties.

I am happy that our paper “The Impact of the EU AI Act’s Transparency Requirements on AI Innovation” has been accepted for presentation at the 19th International Conference on Wirtschaftsinformatik (WI 2024). The conference will take place from September 16 to 19 in Würzburg, Germany.

Companies are faced with the challenge of counteracting the risks associated with the integration of artificial intelligence (AI) in order to maximize the value creation potential of the technology. Traditional governance mechanisms often have to be adapted to cover the specific requirements and risks of AI applications. This article presents a method developed as part of a scientific study that enables a systematic transformation of existing governance mechanisms towards comprehensive AI governance. This method provides a practice-oriented guide that supports governance managers and consultants with concrete action steps to integrate AI-specific considerations into existing governance mechanisms. Through an iterative approach and continuous adaptation to technological developments, the method helps companies to realize the value creation potential of AI while controlling the risks. Please read more on this topic in our (German) paper "Governance von künstlicher Intelligenz – Eine Methode zur Transformation vorhandener Governance-Mechanismen in Unternehmen" recently published in Wirtschaftsinformatik & Management.