Understanding Engineers' Drivers and Impediments for Ethical System Development: The Case of Privacy and Security Engineering

Sarah Spiekermann, Jana Korunovska, Marc LangheinrichWU 2018

TL;DR

This study reveals that while engineers generally recognize the importance of privacy and security in system design, their actual implementation is significantly hindered by organizational factors like weak norms, lack of time, and autonomy. Ultimately, engineers' perceived responsibility is the strongest predictor of ethical engineering practices, more so than traditional behavioral models.

What Problem Does This Paper Address?

As digital systems become increasingly integrated into our lives, ensuring they are developed ethically, particularly concerning privacy and security, is paramount. However, the reality of the business world—characterized by tight deadlines, data markets, and a constant push for perfection—creates a challenging environment for engineers. This paper investigates the drivers and impediments engineers face when trying to incorporate ethical considerations like privacy and security into their work, aiming to understand why these crucial aspects are often overlooked or poorly implemented.

Methods

The researchers employed a mixed-methods approach, combining an empirical study of 124 software and system engineers with six interviews with senior engineers. The empirical study utilized an online questionnaire based on the Theory of Planned Behavior (TPB), which posits that behavior is influenced by attitudes, subjective norms, and perceived behavioral control. Additionally, they incorporated Jonas' Principle of Responsibility to explore the role of perceived personal responsibility. The study measured engineers' reported behavior, attitudes, subjective norms, perceived behavioral control, and perceived responsibility. Statistical analyses, including regression, were used to identify the relationships between these factors and actual engineering practices.

Datasets

Not discussed in the paper. The study involved an empirical survey of 124 engineers and interviews with senior engineers.

Key Insights and Ideas

  • Awareness vs. Action: Engineers largely understand that privacy and security are important (high instrumental attitudes and subjective norms), but this awareness doesn't always translate into action. They often don't enjoy working on these aspects.
  • Organizational Roadblocks: The organizational environment is a major impediment. Weak or conflicting organizational norms, lack of time, and insufficient autonomy significantly hinder engineers' ability to implement privacy and security measures.
  • Perceived Responsibility is King: The study found that engineers' individually perceived responsibility is the most significant predictor of their engagement in privacy and security engineering, even more so than the factors outlined by the Theory of Planned Behavior.
  • "Activist Engineers": In organizations with weak norms, some engineers with high perceived responsibility and control exhibit "activist engineering," pushing for ethical practices despite organizational resistance.
  • Privacy vs. Security Nuances: While both privacy and security are important, the drivers and impediments can differ slightly. For instance, organizational norms play a more direct role in privacy engineering, while a three-way interaction of norms, control, and responsibility is crucial for security engineering.
  • Character Matters: Beyond organizational factors, certain individual traits like a high locus of control, emotional stability, and lower fear of failure are positively associated with ethical engineering practices.

Key Results

  • The TPB models explained 36% of the variance in past privacy engineering and 37% in security engineering.
  • For privacy engineering, only instrumental attitudes and perceived responsibility were significant predictors. Perceived responsibility alone explained 10% of the variance.
  • For security engineering, both instrumental and experiential attitudes were significant predictors, but subjective norm and perceived behavioral control were not.
  • Perceived responsibility was the strongest predictor for both privacy (β = .36) and security (β = .46) engineering, explaining a substantial portion of the variance.
  • 38% of organizations had weak norms for privacy, and 31% had weak norms for security.
  • A significant portion of engineers (around 40%) reported not feeling responsible for privacy and security in their systems.
  • Engineers with high responsibility and high perceived behavioral control implemented privacy and security mechanisms frequently, regardless of organizational norms.

Limitations and Future Work

The authors acknowledge several limitations:

  • Cross-sectional Design: The study's design does not allow for definitive causal interpretations, although the TPB framework suggests causal relationships.
  • Focus on Research Institutions: The sample included engineers from research institutions, which might have different dynamics than purely industry-focused environments.
  • Single-Item Measures: While defended by the authors, the use of single-item measures for some constructs could be a limitation.

Future work suggested includes:

  • Causal Studies: Investigating causal relationships more directly.
  • SDKs and APIs: Exploring the role of Software Development Kits (SDKs) and Application Programming Interfaces (APIs) in shaping engineers' control and ability to implement ethical designs, especially in highly monopolized technical platforms.
  • Deeper Understanding of Responsibility: Further research into what shapes perceived responsibility, whether it's cognitive moral development, education, or organizational climate.
  • Personality Traits: Investigating the interplay of personality traits (locus of control, hope of affiliation, hope of success, agreeableness, fear of failure, religiousness) with responsibility and ethical engineering.

Summary updated

Abstract

Machine ethics is a key challenge in times when digital systems play an increasing role in <br/>people's life. At the core of machine ethics is the handling of personal data and the security of machine <br/>operations. Yet, privacy and security engineering are a challenge in today's business world where personal <br/>data markets, corporate deadlines and a lag of perfectionism frame the context in which engineers need to <br/>work. Besides these organizational and market challenges, each engineer has his or her specific view on the <br/>importance of these values that can foster or inhibit taking them into consideration. We present the results <br/>of an empirical study of 124 engineers based on the Theory of Planned Behavior and Jonas' Principle of <br/>Responsibility to understand the drivers and impediments of ethical system development as far as privacy <br/>and security engineering are concerned. We find that many engineers find the two values important, but do <br/>not enjoy working on them. We also find that many struggle with the organizational environment. They face a <br/>lack of time and autonomy that is necessary for building ethical systems, even at this basic level. <br/>Organizations' privacy and security norms are often too weak or even oppose value-based design, putting <br/>engineers in conflict with their organizations. Our data indicate that it is largely engineers' individually <br/>perceived responsibility as well as a few character traits that make a positive difference.
BibTeX
@inproceedings{spiekermann2018understanding,
  author = {Sarah Spiekermann and Jana Korunovska and Marc Langheinrich},
  title = {{Understanding Engineers' Drivers and Impediments for Ethical System Development: The Case of Privacy and Security Engineering}},
  booktitle = {WU},
  year = {2018},
}
Figure