The purpose of this study is to establish a logical foundation for identifying and addressing bias in AI technology. Bias is not merely an error but an inherent aspect of human cognition and social structures. Rather than attempting to eliminate it entirely, ethical efforts should focus on analyzing and adjusting its effects. AI bias is particularly concerning because it has the potential to reflect or even exacerbate existing social inequalities. Therefore, a careful approach is required at every stage, from data collection and processing to algorithm design. This study examines the logical fallacy underlying AI bias, beginning with the issue of hasty generalization—where conclusions are drawn too broadly based on limited data or specific cases. Just as humans are susceptible to cognitive biases such as confirmation bias and availability bias, AI systems also face similar challenges due to the limitations of their training data. Moreover, bias in AI is often multifaceted. It does not arise from a single factor but rather emerges through the interaction and amplification of various social, technical, and data-related influences. The primary sources of such complex bias include data
📖 افتح في inklap 🔗 DOI 📮 اطلب بحثاً