Ethics of AI: A Systematic Literature Review of Principles and Challenges

Four core principles dominate global AI ethics consensus: Transparency, privacy, accountability, and fairness emerged as the most commonly cited ethical principles across organizations and regulatory bodies worldwide, suggesting these areas should be prioritized in AI development and deployment.

Vague principles undermine real-world implementation: The research identified "vague principles" as a significant challenge preventing organizations from actually adopting AI ethics guidelines, indicating that high-level ethical statements need clearer, more actionable definitions to be effective.

Knowledge gaps are a major barrier: Lack of ethical knowledge was reported as one of the most significant obstacles to implementing AI ethics in practice, suggesting that training and education initiatives are critical for moving ethics from policy documents into actual AI systems.

A comprehensive framework of 22 principles and 15 challenges has been mapped: The systematic review identified a global convergence around 22 ethical principles and 15 specific challenges, providing a structured foundation that policymakers and organizations can use to assess and improve their AI ethics practices.

Maturity models could help organizations improve: The research proposes developing maturity models to evaluate AI systems' ethical capabilities and identify best practices, offering a pathway for organizations to systematically assess and enhance their ethical AI implementation over time.