Invisible Women Exposing Data Bias in a World Designed for Men

The author is a campaigner, writer, broadcaster and activist whose successful campaigns have included getting Florence Nightingale to be the first woman on a Bank of England note and lobbying to get a statue of suffragist Millicent Fawcett in London. Over six chapters she describes how the world is designed by men for use by men. An early comment is particularly relevant for the current interest in the impact of the menopause in the workplace. She suggests that offices are designed with temperatures more suited to male workers than female. I am sure many women would agree. A big feature of the book is the data gap and how that influences everything from design of cars to pharmaceutical drug performance. Right up to date there is consideration of artificial intelligence (AI) and how some algorithms presume all users are men with interesting consequences. Translation software is an example—with female pronouns appearing only half the number of times as male. Image data sets such as Google Images is reported to contain male images vastly more than female and when analysis of those images is undertaken uncomfortable stereotypes emerge. Search for images of ‘CEO’ for example and only 11% of images feature a female CEO. This is despite there being 27% of CEO posts occupied by females (in the USA). Some of the data gaps are frankly dangerous. The author reports how most drugs, including anaesthetics and chemotherapeutics, have gender neutral dosages thus putting women at risk of overdose. Part of this is due to the higher body fat. Which led me to wonder if the same applied to occupational exposure limits—should they be gender based? This is one of those books which after you have read it you begin to see how much work needs to be done and to worry about how the huge data sets currently being used to develop AI applications could go so badly wrong. This book won the Royal Society Science Book Prize 2019 and I can see why. Some of the content is obvious but the way it is pulled together makes a good comprehensive read and an essential text for anyone collecting and analysing data and reviewing and interpreting evidence. ★★★☆ (buy and keep)