If we want the computer systems that run our world to value women, we need more women coders.
In 2016, a New York judge ruled that the state’s education department had discriminated against teachers by using flawed algorithms to evaluate their performance in the classroom. The system, called Value Added Measures (VAM), claimed to be able to measure by how much teachers were improving their pupils’ grades.
There was just one catch, it didn’t always work. And when it didn’t, the consequences for teachers could be severe. In one year, fourth-grade teacher Sheri G. Lederman went from being highly rated to “ineffective”.
It took Lederman four years and an exhausting court case to get to the bottom of what happened. The VAM algorithms, it turned out, didn’t consider things such as the student’s innate abilities in a subject or the size of the class.
If a teacher had a small class, one or two individuals’ scores would have a disproportionate influence on the performance metric. A teacher with very bright pupils who started and ended the year at the top of the marking range, would appear not to have added much value. In both cases, the teacher might unfairly be rated “ineffective”.
This story illustrates a problem inherent in the software development industry. We often code without thinking about our audience, or the people who will benefit from our systems. Imagine if, instead of putting a teacher through a court case to discover the flaws in an algorithm, she had been included in the decision-making process by the people designing and building the system?
More recently, we've heard similar stories about algorithms having a negative impact on people's lives. In 2018, a study by MIT found that artificial intelligence systems designed to recognise human faces had an error rate of 0.8% for light-skinned men but 35% for darker-skinned women.
Researchers at the University of Virginia found that AI-driven image recognition software had been teaching itself to associate pictures of kitchens with women. In 2019, the US financial regulator opened an investigation into Apple’s new credit card, because it appeared to discriminate against women when assigning credit scores.
To avoid these outcomes and improve the quality of software we build so it serves everyone — and all communities — equally well, we need software development teams that are more diverse. A big part of this task is to get more women working in the industry. Right now, only 19% of software developers in the US are women. In the UK that figure is even lower — just 13%.
For this reason, the University of Bath is keen to recruit more women to its online Computer Science MSc and Artificial Intelligence MSc courses. Both courses are taught remotely, via user-friendly digital learning platforms. Applicants for both should have A-Level maths. For the AI course, some coding skills are desirable. But for the online Masters in Computer Science, the only requirement — other than the maths qualification — is that applicants like a challenge, have an enquiring mind, critical thinking skills and an interest in the subject. No previous coding experience is required: we teach it from scratch.
The careers for which these courses qualify you are often extremely well paid. The average salary for a UK software developer is £44, 000, for an AI developer it’s £80,000, and most companies are keen to recruit more women with the necessary qualifications.
Most of the people you work with in software development are interesting, high-functioning problem solvers who prize different perspectives and ways of approaching problems. There are often opportunities to work flexibly and, at least sometimes, from home.
Software development is a creative, challenging field of work that offers those who pursue it the chance to use their insights, experiences, creativity and critical-thinking skills to make a real and human difference to the lives of millions of people.
“The fields of Computer Science and Artificial Intelligence are ever growing. The software we build affects all our lives, yet there are still few women involved in the design and development of these systems. It would be fantastic to see more women join our Computer Science and Artificial Intelligence programmes.”
— Dr. Christina Keating, Computer Science department