AI & technology ethics

Story hook

In 2018, an investigative team at ProPublica in the United States obtained data from a software called COMPAS — used by American courts to predict the recidivism risk of defendants and to recommend bail and sentencing. The journalists examined 7,000 cases in Broward County, Florida. The numbers were chilling: Black defendants were almost twice as likely as white defendants to be falsely flagged as future criminals, while white defendants were almost twice as likely to be wrongly labelled low-risk. No human judge had written racism into the code; the algorithm had learned it from historical sentencing data — and was now amplifying it through every fresh case.

Six years later, in March 2024, the European Parliament passed the EU Artificial Intelligence Act — the world's first comprehensive AI law, categorising systems by risk (unacceptable / high / limited / minimal) and banning real-time biometric identification in public spaces. The same year, Italy temporarily banned ChatGPT over privacy concerns; India's Ministry of Electronics & IT issued an advisory requiring prior permission for unreliable AI deployments (later softened); the US Federal Trade Commission launched investigations into OpenAI, Microsoft, and Amazon. A new branch of ethics — AI ethics — has gone from philosophy seminars to parliamentary statute books in less than a decade.

For UPSC candidates, the question is no longer whether to study AI ethics. It is which of its four pillars — bias, transparency, accountability, algorithmic justice — to deploy in which answer. This unit gives you the vocabulary.

Why this matters for UPSC

AI ethics is the newest GS-IV unit, added in spirit (not syllabus text) post-2020 as questions began appearing — 2020, 2022, 2023 — on algorithmic governance, data ethics, surveillance and digital divides. Prelims has touched it through DPDP Act 2023

• EU AI Act 2024. Mains uses it in case studies (300-word scenarios) and quote-based essays. Interview boards probe it heavily for technical streams (engineering, IT background) and even for arts streams when Aadhaar, face-recognition, or Aarogya Setu comes up.