Conference Agenda

Artificial Intelligence through machine learning becomes increasingly important in civil engineering practice and has been applied, among others, in structural design of building infrastructures, time and cost planning of large projects, and risk quantification. Methods based on machine learning (ML) use large volumes of stored data and identify patterns or relationships within these datasets through a self-learning process. ML technologies “learn” these relationships from training data, without being explicitly programmed.

In essence, the majority of ML technologies describe patterns and real-world phenomenon in a fashion not very comprehensible, intelligible or at the very least rationalizable to human (i.e., black box solutions). Given the ever-increasing accuracy of ML technologies in civil engineering practice, the reliance and dependence of humans on ML-based solutions increase. This can create situations where users of ML technologies perceive this practice from a perspective unbeknownst to them. When introduced in decision-making, final decisions become the result of a complex interplay between designers, users and technology (Fritz, Brandt, Gimpel, & Bayer, 2020; Redaelli, 2022; Verbeek, 2008). A major question is if one can speak of a hybrid agency between these actors. Can one speak of a dialogue between these actors? And under what conditions can AI become smarter than its designer or user? Does AI also learn from its user?

An empirical case study of ML technology for the design optimization process of wind turbine foundations to reduce the overall design time without compromising the accuracy was carried out to deal with these questions. Because in an actual design process, an extensive number of design variables are involved, it is essential to verify and determine the most influential variables. Using ML, the likely influential design variables are determined. But can AI in this use practice become smarter than the design engineer by showing new influential design variables not seen by the engineer? And which actor determines there is a missing influential variable? Does ML provide designers a better understanding of the importance of each design variable and how a certain design variable influences the behavior of the wind turbine foundation?

It is shown that designing and using AI systems in design engineering involve many actors. Because there is a web of responsibilities it is impossible to hold one actor accountable. Concepts from postphenomenology (Verbeek, 2008) may clarify this perceived hybrid agency between users, designers and AI in design engineering practice. By using these concepts, the increasingly close relationship between users, designers and AI in design engineering practice can be examined. In responding to the call of Leiringer and Dainty (2023), applying these concepts can in general have the potential to increase the maturity of civil engineering research.

This paper explores the intricate relationship between socially disruptive technologies and conceptual engineering, emphasizing the necessity for re-evaluating and adjusting traditional concepts in the context of rapid technological advancement. The introduction of the "Conceptual Exportation Question" (CEQ) highlights the complexities of moral judgments within virtual environments, revealing the limitations of existing ethical and metaphysical frameworks in assessing virtual behaviors. The findings indicate that conceptual engineering serves as an effective response to the disruptions caused by technology, offering a novel methodological framework to tackle the diverse levels and forms of conceptual interference.

The research underscores the importance of adapting ethical considerations to the unique challenges posed by emerging technologies, particularly in the realms of information security and virtual ethics. By examining how traditional concepts may falter in the face of technological evolution, the paper advocates for a more dynamic approach to conceptual definitions and applications. Future studies should delve deeper into the implementation of conceptual engineering within fluctuating technological landscapes, focusing on the practical application of CEQ standards. This entails a comprehensive analysis of the variances in concept applicability across different virtual contexts, which is essential for a nuanced understanding of the ethics surrounding virtual actions.

Moreover, the paper argues for the integration of historical and contextual analyses of concepts, drawing on Nietzschean methods to assess the origins and functions of concepts in both non-virtual and virtual environments. Such an approach not only enriches the discourse on virtual ethics but also bridges the gap between traditional ethical considerations and the unique demands of virtual interactions. By recognizing the relative nature of concepts and their ethical implications, the study paves the way for more robust frameworks that can accommodate the complexities of virtual realities.

In conclusion, the research presents a compelling case for the application of conceptual engineering as a vital tool in addressing the challenges posed by disruptive technologies and virtual ethical dilemmas. By fostering a deeper understanding of how concepts can be adapted to meet the demands of evolving technological contexts, the study contributes to the ongoing discourse on ethics in the digital age. It encourages scholars and practitioners alike to consider the implications of technological advancements on moral judgments and to develop more flexible and context-sensitive ethical frameworks that can effectively navigate the intricacies of virtual environments.