Ethics Practices in AI Development: An Empirical Study Across Roles and Regions

Recent advances in AI applications have raised growing concerns about the need for ethical guidelines and regulations to mitigate the risks posed by these technologies. In this paper, we present a mixed-methods survey study - combining statistical and qualitative analyses - to examine the ethical perceptions, practices, and knowledge of individuals involved in various AI development roles. Our survey comprises 414 participants from 43 countries, representing various roles such as AI managers, analysts, developers, quality assurance professionals, and information security and privacy experts. The results reveal varying degrees of familiarity and experience with AI ethics principles, government initiatives, and risk mitigation strategies across roles, regions, and other demographic factors. Our findings underscore the importance of a collaborative, role-sensitive approach that involves diverse stakeholders in ethical decision-making throughout the AI development lifecycle. We advocate for developing tailored, inclusive solutions to address ethical challenges in AI development, and we propose future research directions and educational strategies to promote ethics-aware AI practices.

💡 Research Summary

This paper presents a large‑scale mixed‑methods investigation into how AI development professionals across the globe perceive, understand, and apply AI ethics principles and governance frameworks. The authors recruited 414 participants from 43 countries through platforms such as Prolific, X (formerly Twitter), Reddit, Quora, LinkedIn, Hugging Face, and Kaggle. Respondents represented a wide spectrum of roles within the AI lifecycle: managers (project, product, CEOs, CTOs), analysts, developers, quality‑assurance engineers, researchers, ethicists, and information‑security and privacy specialists.

The survey consisted of three sections: (1) demographic and professional background (education, role, organization size/type, AI experience), (2) general perceptions and practices related to AI, and (3) detailed knowledge, experiences, and risk‑mitigation strategies concerning AI ethics principles, regulatory initiatives, and best practices. The authors formulated four research questions (RQs).

RQ1 examined overall AI perceptions. Participants largely associate AI with process automation and performance enhancement. Even non‑AI specialists use AI daily for productivity gains, while data‑protection and security concerns remain salient among those who have not yet integrated AI into their workflows.

RQ2 explored familiarity with AI ethics principles and governance initiatives. Familiarity varied significantly by role, organization size, gender, and geographic region. Oversight roles (product managers, requirements analysts), researchers, employees in government or mid‑sized firms, and female participants reported the highest awareness of most ethics principles. Intra‑Class Correlation analysis revealed greater variability within roles than between roles, suggesting that individual experience and team culture drive ethical awareness more than the role label itself. A LASSO regression identified knowledge of the EU AI Act as the strongest predictor of overall ethics‑principle familiarity, underscoring the impact of concrete regulatory education.

RQ3 investigated how ethical principles are operationalized across roles. Managers and requirements analysts tend to embed ethics guidelines into documentation and place strong emphasis on user rights. Academics and researchers focus on technical dimensions such as transparency, explainability, fairness, and justice. Mixed‑effects modeling confirmed “role” as the primary predictor of ethical practice, with organization size and industry sector providing secondary influence. Qualitative comments highlighted that government agencies and mid‑sized companies often possess more robust ethical frameworks, whereas startups and smaller teams rely on informal practices. Gender differences emerged, with women generally assigning greater weight to ethical considerations throughout development.

RQ4 examined risk‑perception and mitigation strategies. Administrative roles (product managers) prioritize establishing clear ethical policies, delivering ethics training, and communicating principles across teams, frequently leading risk‑impact assessments. Developers employ complementary tactics such as evaluating model outputs across diverse populations to curb bias. Security and privacy specialists focus on traditional safeguards (encryption, access control) rather than broader AI‑ethics concerns. Researchers depend on institutional policies or personal judgment when developing AI tools, reporting challenges in drafting ethics statements due to ambiguous guidance and the difficulty of balancing multiple ethical dimensions.

Across all RQs, the study demonstrates that ethical awareness and practice are not uniform; they are shaped by a complex interplay of professional role, organizational context, gender, and regional regulatory exposure. The authors argue for a role‑sensitive, collaborative approach to AI ethics that integrates tailored education, clear governance, and cross‑functional communication. They propose future work on formalizing role‑based ethical decision‑making processes, developing maturity models for responsible AI, and creating practical toolkits that translate high‑level principles into actionable development‑stage activities.

In sum, the paper provides the first comprehensive, multi‑regional, multi‑role empirical assessment of AI ethics practices, revealing critical gaps and offering concrete directions for academia, industry, and policymakers to foster ethically aware AI development.