Encoded Inequality

AI and ML are more innovative due to their capacity to process vast data sets, automate intricate decision-making processes, and foster creativity across many industries. These features will enhance productivity and efficiency. The data utilised to train AI systems and the algorithms governing them are biased, thereby limiting their potential. This chapter analyses the biases in AI systems stemming from data collection methods, algorithmic biases, and their societal impacts. Information can exhibit bias on demographics and societal prejudices that AI systems inadvertently perpetuate. Such biases may cause AI systems to preferentially favour or unjustly discriminate against specific groups, resulting in inequity. This problem is examined through case studies and experimental research utilising advanced AI models, such as DALL-E 3. These approaches facilitate the examination of AI bias discovery and mapping both independently and within its contextual framework. The chapter also suggests measures to mitigate these biases, promoting fairness and openness in AI systems.