Racing to the Bottom: Impact of the Global AI Race on Ethical and Responsible AI Development

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

As one of the most pressing influences over the future course of AI development in the twenty-first century, the AI Race can be defined as a contest among nations for dominance in the domain of AI research and development. Rooted in geopolitical competition among world powers, increased business competition among firms, and an iterative and provisionally adaptive culture, the AI Race introduces systematic incentives that oppose responsible AI. This essay analyzes the nature of the AI Race and its effects on responsible AI, with special reference to the eight AI Ethics Principles proposed by the Department of Industry, Science and Resources in Australia [4]; UNESCO’s global Recommendation on the Ethics of Artificial Intelligence [10]; and Hagendorff’s [6] meta-analysis of AI ethics guidelines. Specifically, issues of importance considered in this study include: the degradation of safety and reliability under competitive pressure timelines; the deepening problem of bias and unfairness due to insufficiently audited AI development pipelines; increasing AI opaqueness resistant to human monitoring and control; decreased privacy due to competitive pressure for data acquisition; and the expanding chasm between the capabilities of rapidly evolving AI and regulation governing its use. The study draws upon a structured review of authoritative and peer-reviewed primary sources and secondary literature published up until 2026, thus offering analysis at a crucial moment of inflection within the AI Race. It is argued that the AI Race represents a multidimensional threat to responsible AI, which cannot be sufficiently addressed merely via the implementation of ethical principles. Although the eight AI Ethics Principles offered by Australia are a substantive set of provisions, compatible with similar guidelines worldwide, Hagendorff’s [6] analysis suggests that voluntary approaches fail to generate any changes in industry. Moreover, the findings of Deloitte’s [3] survey demonstrate that organizations in Australia fall behind their global peers in terms of AI governance maturity, while Darden’s [7] analysis predicts a narrow window of opportunity over the next five years for building ethical infrastructure in AI before systems are integrated into essential social/governmental functions. Five practical recommendations are made based on the analysis: introducing a mandatory tiered AI risk system (similar to the EU AI Act); tying government AI procurements to compliance with Australia’s ethics principles; promoting Australia’s role as a leading player in international AI standards setting; investing in AI governance capacity within Australia; and considering a paradigm shift in thinking about ethical AI as critical national infrastructure.