Special Issue on "Trust in AI for Electronic Markets"

Call for Papers: Trust in AI for Electronic Markets

Guest Editors

Wolfgang Maass, Saarland University and German Research Center for Artificial Intelligence (DFKI), Germany, wolfgang.maass(at)dfki.de

Roman Lukyanenko, HEC Montréal, Canada, roman.lukyanenko(at)hec.ca

Veda C. Storey, Georgia State University, USA, vstorey@gsu.edu

Theme

Electronic markets for trading physical, as well as digital, goods offer a wide variety of services based on Artificial Intelligence technologies, as smart market services. Smart market services generate recommendations and predictions using Artificial Intelligence (AI) technologies on data available and accessible in electronic markets. For instance, financial high-speed trading is only feasible by smart market services that autonomously execute transactions according to market signals based on AI models trained with big data. Electronic marketplaces, including Amazon and Alibaba, are using AI technologies to provide smart services to consumers, optimize logistics, analyze consumer behavior, and derive innovative product and service designs. Some business leaders even consider there to be major threats to society from sophisticated AI solutions, while using AI extensively for their own business. Because AI systems elude human understanding and scrutinization, trust in AI is crucial for the success of smart market services, as well as other AI or machine learning-based systems . Gaining trust in AI begins with transparency in the reviews of (a) data so that biases and gaps in knowledge of a domain are controlled, (b) AI models and objective functions, (c) model performance and (d) results generated by AI models for decision making. Trust becomes an important factor for overcoming uncertainty on AI-based recommendations in general and in electronic markets in particular.

The quality of smart market services depends on shared understanding and conceptual models of data used for training AI models; data quality; the selection and training of appropriate models; and the embedding of models into smart market services. Providers of smart market services are required to build trust relationships with business and end customers based on limited possibilities for opening the “black boxes” of Artificial Intelligence systems due to increased complexity of machine learning models. Empirical studies on trust in AI indicate heterogenous results. Companies and end-users appreciate benefits and opportunities provided by smart market services. At the same time, concerns are raised with respect to privacy issues and biases of data, models and algorithms. Overly optimistic customers might become disappointed if smart market services do not deliver as expected. Proof of privacy leaks and biases might reinforce prejudices. Both may lead to decrease of trust in AI. Challenging research questions are to identify which methods, indicators and experiences have increasing effects on trust in AI. For instance, explainable AI is a technical means for opening “black boxes” of AI systems, generally, and smart market services, specifically.

This special issue seeks contributions on trust in Artificial Intelligence in the context of electronic markets. Contributions that help to understand challenges from an economic, legal or technical perspective are invited.

Central issues and topics

Possible topics of submissions include, but are not limited to:

  • Trust behavior and AI
  • Mental models, conceptual models and AI models
  • Psychological and sociological factors for trust in AI
  • Human-centric design of smart market services
  • Explainable AI for smart market services
  • Threats for trust in AI
  • Frameworks for smart markets
  • Business and legal aspects influencing trust in AI
  • Relationships between trust and Business models with smart market services
  • Transparency of data, AI models and recommendations
  • Case studies on building trust in AI

Submission:

Electronic Markets is a Social Science Citation Index (SSCI)-listed journal (IF 2.981 in 2019) in the area of information systems. We encourage original contributions with a broad range of methodological approaches, including conceptual, qualitative and quantitative research. Please also consider position papers and case studies for this special issue. All papers should fit the journal scope (for more information, see www.electronicmarkets.org/about-em/scope/) and will undergo a double-blind peer-review process. Submissions must be made via the journal’s submission system and comply with the journal's formatting standards. The preferred average article length is approximately 8,000 words, excluding references. If you would like to discuss any aspect of this special issue, you may either contact the guest editors or the Editorial Office.

Keywords:

Trust, Interpretability, Mental Models, Conceptual Models, Explainable AI, Smart Market Services, Privacy, Fairness of Artificial Intelligence, Biases, Transparency

Important deadline

* Submission Deadline: December, 15, 2021