An exploratory qualitative analysis of AI ethics guidelines

Abstract

Purpose As Big Data and Artificial Intelligence (AI) proliferate, calls have emerged for ethical reflection. Ethics guidelines have played a central role in this respect. While quantitative research on the ethics guidelines of AI/Big Data has been undertaken, there has been a dearth of systematic qualitative analyses of these documents. Design/methodology/approach Aiming to address this research gap, this paper analyses 70 international ethics guidelines documents from academia, NGOs and the corporate realm, published between 2017 and 2020. Findings The article presents four key findings: existing ethics guidelines (1) promote a broad spectrum of values; (2) focus principally on AI, followed by (Big) Data and algorithms; (3) do not adequately define the term “ethics” and related terms; and (4) have most frequent recourse to the values of “transparency,” “privacy,” and “security.” Based on these findings, the article argues that the guidelines corpus exhibits discernible utilitarian tendencies; guidelines would benefit from greater reflexivity with respect to their ethical framework; and virtue ethical approaches have a valuable contribution to make to the process of guidelines development. Originality/value The paper provides qualitative insights into the ethical discourse surrounding AI guidelines, as well as a concise overview of different types of operative translations of theoretical ethical concepts vis-à-vis the sphere of AI. These may prove beneficial for (applied) ethicists, developers and regulators who understand these guidelines as policy.

Ethical AI in practice: Balancing technological advancements with human values

“…The study highlights the diversity of values promoted in these guidelines and underscores the need for greater reflexivity in their ethical frameworks. This analysis provides valuable insights into the ethical discourse surrounding AI, offering a nuanced understanding of how ethical principles are operationalized in various AI contexts (Franzke, 2022). The qualitative approach adopted in this study, supported by the theoretical frameworks discussed, provides a comprehensive understanding of the ethical landscape in AI.…”

Bridging Legal Methodology and Ethical Considerations: A Novel Approach Applied to Challenges of Data Harvesting

“…Research ethics, particularly in medical and psychological research, have historically set the standards for ethical considerations in other fields. As a practical discipline of philosophy (van den Hoven, 2010), ethics first addressed areas where life and health were directly affected, taking on both a guiding and regulatory role (Franzke, 2022). Relevant ethical codes, such as the Declarations of Helsinki (World Medical Association, 2022) and Taipei (World Medical Association, 2020), as well as the research rules formulated by the American Psychological Association (2017), have established standards for informed consent and the absence of deception.…”

“…The ethical discussion surrounding data harvesting has evolved to address issues beyond informed consent. The aforementioned meta-reviews by Jobin and Ienca (2019), Franzke (2022), and Holtorf et al (2023) reveal a focus in ethical codes on identifying and balancing competing values.…”

“…The rise of data harvesting has highlighted the need for ethical guidance and legal regulation, as it brings to the forefront a range of ethical concerns, from privacy issues to the handling of digital identities. Numerous papers, as summarized in the meta-reviews by Jobin and Ienca (2019), Franzke (2022), and Holtorf et al (2023), have contributed to a growing body of literature addressing these ethical challenges. In parallel, there has been an increase in legal regulation, particularly in the European Union (EU), notably through initiatives such as the General Data Protection Regulation (GDPR) and the proposed Artificial Intelligence Act.…”

Key Considerations of Ethical Artificial Intelligence That Organisations Need to Consider for Success

“…For instance, the article reviewed 70 international ethical guidelines and presented four key areas common in existing guidelines. These are values, big data and algorithms, inadequate understanding of AI ethics and values related to transparency and data security (Franzke, 2022). From a corporate perspective, IBM (2022) presents three key insights in their report.…”