No surprise here: AI will hit women and low-paid workers hardest and first

Women and those with lower-grade jobs will be significantly more affected by AI automation, according to a report from the Institute for Public Policy Research (IPPR). To which I say: no kidding.

The IPPR think tank decided to investigate which roles were most at threat from generative AI — and wisely also considered who tends to fill those jobs — so examined 22,000 tasks in the UK economy, across all jobs. 

The report split AI adoption into two stages: the first wave, which is now; and the second, which will see AI more deeply integrated into businesses. In that first wave, 11% of tasks are already being impacted, from routine cognitive tasks such as database management to organisational tasks such as scheduling, the report says.

Many of those tasks are in back-office, entry-level and part-time jobs, such as secretarial and customer service roles, where women make up much of the workforce. Younger people are also in trouble, as companies may automate entry-level jobs, giving them no foot in the door.

But when the second wave of AI hits, as many as 59% of tasks could be impacted, the report claims. And that could cost the UK as many as eight million jobs, the think tank predicts, as the impact hits higher paying “non-routine white collar tasks” such as taxation experts, graphic designers and finance brokers.

Further down the line, the use of avatars could transform teaching, medicine and hospitality, though the think tank admits that will require a serious societal shift in what we’re willing to accept from our GPs.

Now, to be clear, having a task affected by AI is not necessarily a bad thing, as it could automate the dullest of rote tasks leaving more time for creative work — that’s the optimistic view. But the IPPR believes there will be outright job losses, too.

AI’s impact on women and customer service roles

To start, let’s set aside AI and examine existing societal structures.

Women are more likely to be replaced by AI because they do the bulk of low-level administrative work. This is what will be done away with in this first wave of automation (assuming IPPR is correct).

Perhaps we should address that. Perhaps we should make it easier for women to smash the glass ceiling, access higher-paid jobs in “traditionally male” industries (in particular IT), and offer enough flexibility to allow time to care for children and older parents without losing their career trajectory.

My point is this: if we had more male customer service reps and male back-office staff, then this report wouldn’t be about gender, just job losses. 

Secondly, let’s reconsider what actually happens in those roles.

Customer service is often seen as easily automated. But tell that to customers tired of pressing phone buttons or coming up against digital walls with chatbots. Indeed, tell that to Air Canada, which employed an AI-powered chatbot to answer customer queries, only for the bot to give bad advice which cost the customer; a lawsuit ensued and the airline lost.

Customer service reps don’t just regurgitate corporate policies. They are a company’s front line for sales, marketing and lawsuit-avoidance. Some basic or straightforward customer-service questions can be answered by AI. But others require (and always will) a human touch.

AI should be deployed to give customer service reps more time to offer real customer service — and cut wait times — rather than slash jobs and boost corporate profits. This will be clear after the first round of job cuts, and companies will start rehiring.

Hidden impact of slashing jobs

Slashing little-valued jobs will have knock-on impacts. Consider those entry-level jobs.

If the bottom rung of the ladder is removed for graduate students, that’s bad news for companies — where will they find staff to top up their numbers when needed? Clearly, younger staff need to start somewhere and will merely need help — and perhaps further in-house training — to leap onto the next-from-bottom rung.

Not persuaded? How about this. Mid-level managers — so often male, and always better paid than front-line staff — may think AI isn’t coming for them, but they are incorrect. If you’ve no-one left to manage, what do you think happens to your job?

And if the jobless masses can’t afford your product, even the CEO loses his job. Disruption doesn’t move in one direction, after all.

My point is this: a job is more than the sum of its tasks. And it will soon become clear that we do need people in these roles, even if some of the tasks are automated away.

What we do about AI coming for our jobs

To be clear, the IPPR report is right to note that we need to consider the negative outcomes of this potential technological upheaval as soon as possible. And it’s right to acknowledge that it may well be the worst off in our society who are hit the hardest.

However, it needs to realise that this is usually the case. It’s a good reason to call for wider diversity among the companies developing these technologies and the governments regulating them.

That’s not among the ideas the IPPR says could help stave off the worst impacts of AI disruption. Instead, it calls for a job-centric industrial strategy for AI that would protect existing jobs, ensure gains for workers (not just Silicon Valley) and to support job transitions through retraining programmes.

It also says government should address the fallout from lower labour demand, be it through AI or wealth taxes and a stronger social safety net, including ideas like Universal Basic Income.

All of those are good suggestions — for now, not if and when the AI apocalypse hits.

Related: The biggest challenges faced by Women in Tech today

Not just about AI

Regardless of whether or not AI is part of my role, shouldn’t the work I do benefit myself and not just executives and shareholders?

The pay gap between staff and CEO has leapt 1,460% since 1978, worsened when you remember the gender pay gap still exists, too. That means a stronger social safety net is required, and a wealth or corporate tax is one way to fund it. You can disagree with the politics and the policies all you want, but there’s no question that such challenges exist.

Then there’s retraining.

More than half of UK workers, according to the Office for National Statistics, are professionals, administrators, or the like. But what we need is more people who don’t work in an office, be it in trades or social care. If these latter jobs were better respected and remunerated, the rise of AI would have less of an impact — we already need more people going into these jobs as it stands, so losing a few back-office positions and helping retrain workers to fill desperately needed roles like social care makes sense.

That training not only needs to be in place immediately (yesterday, frankly) but we need to reconsider what these jobs are paid. And pay needs to be higher to smooth the disruption from the AI revolution (if it happens) but also to draw people out of work with dwindling opportunities to where we need them.

As it stands, women will simply be forced to retrain for a lower paid, harder-to-do job.

AI disruption may well be looming, but all of these problems already exist in our economy today. If it takes the fear of ChatGPT to fix how we employ women and what we pay front-line workers, so be it — but it needs doing either way.