Detail

AI can improve workplace efficiency but also reinforce or amplify human biases if not carefully designed and monitored. Amazon's failed AI recruitment tool, which was trained on historical hiring data and ended up discriminating against women by favouring characteristics associated with male applicants. The case became a landmark example of the importance of responsible AI adoption. Experts stressed that human oversight remains essential. Ms Emma Meyer, Senior Manager at SP’s Business Innovation Centre, said AI systems trained on historical data can inherit past biases, potentially affecting recruitment, performance evaluations, and career progression. She emphasised that organisations need high-quality training data and should recognise that AI can amplify biases at scale if left unchecked. AI can be misused if deliberately programmed with discriminatory rules, citing a case involving online education company iTutorGroup, which was found to have automatically rejected older job applicants in the United States. In addition, the experts cautioned that AI-powered workplace monitoring systems and AI-driven performance evaluations should be used carefully, as an overreliance on measurable data could overlook important qualities such as collaboration, leadership and workplace context.

[The Straits Times]