A few years ago, Amazon built an AI tool to screen résumés. It learned from a decade of hiring data — mostly men — and began to downgrade any CV that carried the word 'women.'

No one set out to build a biased machine. That is the point: when women aren’t in the room where these systems are trained, the bias just ships. Last year, we published Wired for Impact: Women in Ind(AI) report to put numbers to that test. Months on, little has changed. Women are still 43% of India’s STEM graduates and barely 20% of its AI operators. The funnel narrows as you climb. That’s the uncomfortable part — the technology has raced ahead while the representation building it has stood still.

A system built by half the population can only ever see half the picture. Which is why inclusion in AI was never about fairness alone. It’s about accuracy.

- Vani Kola

Click here to read the full newsletter