Source: Ikigai Law

The article by Ikigai Law summarizes a roundtable discussion on the challenges of bias in artificial intelligence (AI) systems, held in New Delhi. The event gathered 25 experts, including academics, policymakers, startup founders, and researchers, to propose actionable solutions. Participants emphasized that bias is inherent in AI due to the societal and cultural nature of its training data. Key insights included the impossibility of creating bias-free AI, the importance of context-specific evaluation, and the need to shift focus toward managing socially undesirable biases rather than eliminating all biases. The roundtable proposed solutions such as involving local communities in identifying biases, developing culturally aware evaluation frameworks, and prioritizing open-source AI models for transparency.

Discussions also highlighted the distinction between bias and fairness, stressing that fairness involves aligning biases with ethical norms while mitigating harmful or discriminatory outcomes. Contextual assessment was deemed crucial for understanding biases in specific AI applications. Participants noted the limited resources for low-resource languages like Indic languages and suggested integrating diverse, high-quality multilingual data. They also advocated for flexible governance models, international collaboration, and increased AI literacy for users. Ultimately, the event emphasized the need for adaptive, inclusive, and culturally sensitive approaches to develop AI systems that reflect global diversity and fairness.