According to the 2019 World Trade Report, service trades are likely to increase their share of global trade by 50 percent until 2040. Services will benefit most likely from increasing the automatization and digitalization of former face-to-face processes, and from an increasing demand of online services due to demographic change. The WTO states that global cooperation has to be increased such that all economies can collectively benefit from increasing service trade.

With the globalization of services comes a globalization of knowledge. According to the Research Perspectives of the Max Planck Society, globalization is a nonlinear process, which can lead not only to homogeneity and the standardization of culture, but also to an increase in complexity, as tools and ideas tend to outpace cultural progress. As face-to-face problem-solving will be replaced more and more by digital services, global problems that require global cooperation will have to gain competence in global and complex problem-solving (CPS).

A prominent example of such a complex, global problem is anthropogenic climate change. The Intergovernmental Panel on Climate Change (IPCC) challenges the high imponderability of climate change and its impact on decision-making and policies with their “Integrated Risk and Uncertainty Assessment of Climate Change Response Policies”. In their report, the IPCC states the understanding that decision-makers tend to rather base their decisions on intuitive thinking processes than on thorough analysis and that the perception of risk has to be included in climate change risk management (Kunreuther et al., 2014).

Human decision makers are led not only by rational decision-making, but insights derived from behavioral economics show that people are guided by intrinsic motives, bias, and myopic interpretations of feedback—casting doubt on whether humanity is capable of effectively solving complex problems of global proportions.

With growing successes in the area of artificial intelligence (AI), the United Nations Economic and Social Council has stated concerns that AI may not only offer advantages, but also

“disrupt societies in fundamental ways”,

with people being replaced by automated decision-making devices (United Nations Economic and Social Council, 2019, p. 5). For this reason, talent search is of crucial importance to support domains threatened to be replaced by artificial systems. The UNO High-level Committee on Management places a focus on the identification of talent by automated processes in the area of assessment and testing (United Nations Economic and Social Council, 2019). The hybrid approach of embedding expert knowledge into neural networks, commonly used for AI systems, has been suggested and implemented through the combined effort of various institutes (Barca, Porcu, Bruno, & Passarella, 2017; Chattha et al., 2012; Silva & Gombolay, 2019), raising questions regarding the accountability and regulation of such AI-guided decisions (Doshi-velez & Kortz, 2017). Since the global-employment-changing economic crisis in 2008, the creation of sustainable employment has become a core goal for European institutions, such as the European Foundation for the Improvement of Living and Working Conditions. For systems to act sustainably, they must be flexible and resilient, while knowledge about a system’s state is key (Jeschke & Mahnke, 2013). The European Commission further increased flexibility of the European “Stability and Growth Pact” in 2015, to “build up fiscal buffers” for its member states; these buffers were indeed implemented successfully, according to a 2018 report by the European Commission (European Commission, 2018).

The search for expert knowledge is guided not only by ethics. In trying to gain knowledge of a system as large and complex as the European market, obtaining sufficient amounts of empirical data can be a challenge. Expert knowledge can be used to replace missing data in order to support sound predictions. With highly complex problems comes uncertainty, especially when empirical data is limited. Psychological observations have shown that expert knowledge tends to be biased, when expert knowledge faces uncertainty unguided (European Food and Safety Authority, 2014). Expert identification and management have been suggested by the European Food Safety Authority to be organized in a structural manner, and should result in a database of experts. The “Division for Sustainable Development” of the United Nations Department of Economic and Social Affairs builds upon multi-agent action networks, consisting of resources, knowledge and experts in order to achieve their global sustainability goals. In their 2016 report, the top three challenges listed as related to such networks are limited financial resources, followed secondly by changing mindsets and change management, and thirdly, by human resources, as depicted in figure 1.1 (Division for Sustainable Development. United Nations Department of Economic and Social Affairs, 2016, p. 13).

Future economies will inevitably face global problems, due to the ever-growing connectivity and service-oriented trade. Novel ideas and technological breakthroughs will outpace slow cultural development leading to increasing complexity. Global asymmetries in knowledge and information will further fuel change, making routine problem-solving unreliable and making its outcomes volatile, thus endangering those who cannot maintain modern workspace requirements.

Figure 1.1
figure 1

Source Division for Sustainable Development. United Nations Department of Economic and Social Affairs, 2016, p. 13

Top challenges of modern decision-making networks.

Full size image

CPS and non-routine decision-making experts need to be identified and placed in an environment, where their actions are the most fruitful, such that others can imitate and learn from their success. This scenario could be enabled by a cheap and effective online assessment tool, as financial resources are limited by default. As expert knowledge is especially biased when addressing problems under uncertainty, this thesis focuses on two major goals: i) development of a non-routine problem-solving (NPS) assessment in the form of a highly efficient, online, web browser-based software tool; and ii) obtaining empirical results related to the human individual and group decision-making (GDM), faced with uncertainty, change, different system states, and various forms of environmental public information.

Data and according information on human decision-making behavior was acquired by a randomized experiment, which is considered to be the “gold standard” in scientific research (Rubin, 2008). As in any experimental design, participants were randomly assigned to different public information conditions, where circumstances were actively manipulated. The experiment was both run off- and online, however, the online experiment granted many advantages over its offline counterpart, mostly being more cost-efficient, and enabling the possibility to model all participants’ perspectives via strategy- or logic-categories. Experiments are considered to increase innovation (Kohavi, Longbotham, Sommerfield, & Henne, 2009) and cost-efficient online assessments may support institutions and companies alike in finding experts, assigning them to their most skill-effective working domain, measuring and controlling the impact of information and ultimately supporting management in coping with complex problems successfully.