Artificial intelligence is often seen as a tool for efficiency and progress, but the Harvard Kennedy School paper "How AI Fails Us" (correctly) highlights its limitations and potential harms. The report explores how AI systems frequently reinforce biases, disproportionately affecting marginalized groups in areas like hiring, policing, and healthcare. Structural flaws in data and governance mean these failures are not just technical but systemic, reflecting broader social inequalities. The paper also critiques the lack of accountability in AI deployment, arguing that companies and regulators have struggled to prevent harm. In addition, it also discusses the potential effect of AI on the labour market and workforce. These issues raise questions about whether AI, as currently designed, serves society equitably or amplifies existing disparities.

Reflecting on this, consider your own interactions with AI. Have you encountered an algorithm that seemed biased, misleading, or ineffective? If not, think about cases you’ve read where AI failed—using news online, newspaper or business cases you heard about.

Who is responsible for these failures, and what should be done to mitigate them? Should AI development prioritize transparency, fairness, or oversight? In your reflection, critically analyze AI’s role in society and propose ways it could be improved to ensure it benefits everyone, rather than reinforcing existing power structures.