Whenever I discuss artificial intelligence with people, if they're not envisioning an Ultron-style apocalyptic future, their first question tends to be:

"Are we about to go through a second industrial revolution? Will I be pushed out of my job?"

Countless studies have shown that AI indeed has the potential to displace a part of the workforce, and most have focused on the type of occupation at risk of being displaced. But few discuss the impact AI will have on working women: on the one hand, how would massive job displacement impact a portion of the population that wasn't a part of the first industrial revolution? On the other hand, given the fact that machine learning optimizes processes by predicting the next decision based on a wealth of past actions, if AI starts playing a part in recruitment processes, how do we ensure it doesn't repeat the human biases of racism and sexism which are unfortunately still at play today?

It is no secret that gender stereotypes affect the way women are perceived in the workplace, and the types of roles society pushes them towards. According to a 2013 ONS report, 82% of UK jobs in the ‘caring and leisure’ industry are staffed by women, 77% for ‘admin and secretarial’, and 63% of sales and customer service’.

If you’re as obsessed with AI and feminism as me, you will have noticed two things: first of all, these jobs are people-oriented (ahem gender stereotypes). Second of all, two out of three of these categories are the ones most at risk of displacement by AI: digital assistants, customer next best event prediction systems and other, slightly more obscure systems, are here to reduce the amount of people involved in sales and customer service, whilst Attended Desktop Automation systems reduce the manual workload of admin, in effect also reducing the amount of people needed for the job.

Given the fact that the same ONS report also states that 37 per cent of men are employed in “upper-middle skilled roles”, compared with 18 per cent of women, this puts the women at quite high risk of displacement.

On the other hand, as society grows older and manual tasks become automated, some organisms predict the workforce will shift towards human-oriented jobs: caring and leisure, typically jobs which wouldn’t be automated due to the highly personal, emotional nature of the jobs, will grow, presumably attracting more women workers.

That isn’t necessarily a good thing: the way I see it, we shouldn’t be banking on our inherently sexist thought process to save women by pushing them in even more stereotypical jobs (and where they’re also less likely to get promoted - but that’s a topic for a whole other article). What we should be doing instead in removing barriers to entry towards typically male-dominated fields, ensuring women don’t get pushed out of the workforce by a combination of automation and sexism.

This is also linked to the second way in which women should fear AI: if recruiters start relying on predictions to hire personnel, the strong male-centric bias in past decisions will accentuate and women are at risk of never seeing the jobs they apply for.

Of course, the solution there is quite obvious: don’t use AI for recruiting. If what you aim to do is recruit a diverse range of people, then basing your future choices on past ones will never achieve your goal.

There is more than just the gender they announce differentiating the CV of a man from that of a woman: because women are at risk of being called ‘bossy’ or ‘aggressive’ when acting the same way as their male counterparts, the way in which they market themselves differs.

How do we make sure, then, that women don’t get pushed out of the workforce by a combination of robots and ingrained sexism? Well, first of all, we keep pushing for quotas: if you are forced to recruit women, and a diverse set of them too, then you are ensuring role models for a next generation where, hopefully, women aren’t pushed towards gender-stereotypical jobs.

At an earlier stage, though, we need to educate our girls to teach them they are just as good as boys. According to a recent article in the FT, the change happens between the ages of 5 and 6, when girls start interacting with their environment and internalising the stereotypes they observe.

A good way to start would be to stop gendering our toys so much: whilst I loved playing with Barbies when I was younger, I was also lucky enough to have parents who gave us Legos and G.I. Joes to play with. Rather than impressing on girls the idea that, when they’re playing in a kitchen, they’re playing at being a woman, let’s give them the opportunity to decide whether they would rather build a car.

What this all comes down to may be accelerated by AI, but is a problem we’ve been trying to solve all along: we need to make sure women are well represented across the board, rather than concentrated in stereotypical industries, and we need to promote them, to hire them, to educate them. Let’s give our girls a choice to build whatever they want to build, to give them to opportunity to become whoever they want to become.

#Feminism #ArtificialIntelligence #DailyDilemma