Data

For a more comprehensive exploration of this topic, please read Invisible Women by Caroline Criardo Perez. 

Read any of the articles on our website and you will soon discover that the world is generally built for men. This is proved by the irritating sizes of pockets or phones or the life-threatening testing of drugs and cars. The omission of women is not just biological, but societal and systemic. Women’s concerns are forgotten in city planning, their mental health is forgotten in parenting laws, and their free will is forgotten when discussing abortion. As the seminal book by Caroline Criardo Perez puts it: women are invisible.

The terms “gender data scarcity” and “gender data gap” refer to the phenomenon that women and girls are simply missing from databases. Put in real-life terms, the United Nations (UN) Sustainable Development Goal No.5, namely to “achieve gender equality and empower all women and girls,” is genuinely impossible to achieve. How can that be? Essentially the UN does not know what needs to be done to achieve this goal because there is not enough data available on the difference in needs between women and men. Until now, all policies have been necessarily designed inside this data gap: it is impossible to create gender-informed policies without any data to inform you. 

Let’s zoom in for a moment. Artificial Intelligence (AI) has revolutionised the way the world works. To create intelligent machines that can think and act like humans, AI systems learn from data and can be trained to make decisions and take actions based on the information they receive. I think you can see where this is going. Bias in AI occurs when the data used to train the system is unbalanced or represents a skewed sample of the population. This results in AI systems that are unfair or discriminatory towards certain groups of people. This includes all minoritised groups and also women. 

AI systems have no choice but to exhibit gender stereotypes and discriminate against individuals based on their gender. If the data used to train the AI is not representative of the population, then the AI will be biased towards the data it has the most of, namely white, able-bodied, upper-class, cis-heterosexual men. The AI will make decisions biased towards men and will replicate the actions and thought patterns of men, favouring men in any circumstances. For instance, AI is increasingly used in the recruitment process. 

Back to data on a larger scale: what can we do to fix this problem? The data sceptics out there may very well think women have a one-up on men, and may even reject the problematic binary assumption that the data gap presents. Whilst I agree that the solution to the data gap can not be to create a new gap by excluding those who do not fit into the categories ‘male’ and ‘female’, I do believe we must move towards closing the blanket exclusions, which currently exist. Investing in the collection of national statistics that record the lived experiences of non-men, and addressing biases in classifications, definitions, and even methodologies is essential to move towards a fairer world. 

Identifying inequalities between men and women is often challenging as it perpetuates the very binary that, as a genderqueer person, I am fighting against. But in regards to closing the data gap, I always view more diverse data as a win. Minoritised groups will benefit so massively, perhaps even have their lives saved by the reduction of assumptions in healthcare and safety, but that’s before we even think about how governments manage resource scarcity, also governed by exclusionary data. Whilst working on a binary is never the answer, the World Economic Forum estimates that the data gap is closing so slowly it will take 132 years to reach full parity. People are dying. This issue might not be sexy or eye-catching but it underlies the whole problem of our society.