AI systems can inadvertently perpetuate biases present in the data they are trained on. Leaders must ensure that their AI systems are trained on diverse and representative datasets. Regularly auditing AI outputs for bias and implementing corrective measures are also essential. Promoting diversity within AI development teams can further help mitigate biases.