Every time we make two steps forward with modern technology, it seems like we inevitably end up taking one step back.
Digital bias is the epitome of this issue. Just think about the amazing inventions and innovations we’ve been able to develop in recent years: facial recognition technology, machine learning techniques, artificial intelligence. They’re all incredible and revolutionary in their own way… but they’re also all, unfortunately, biased.
In 2019, the New York Times published a piece titled, ‘Facial Recognition’s Many Controversies, From Stadium Surveillance to Racist Software’. In the article, they discuss webcams that are unable to recognise black people, photo recognition software that incorrectly identifies darker skinned individuals as “gorillas”, and Amazon technology that failed to recognise women (especially those with darker skin) up to 31% of the time. And these are not isolated incidents.
Time and time again, technology is proven to be biased against women and minorities – but it’s not the tech itself that’s the problem. If we dig a little deeper, we find that technology is more than capable of acting in an unbiased manner; the problem is that the researchers behind them are not.
In other words, tech’s favouritism of certain demographics (and, by default, exclusion of others) has led to a lack of diversity, which in turn has led to machines that do not understand much outside of the ‘norms’ they have been programmed with.
History of bias
Unfortunately, this problem isn’t something we’ve only just discovered. It’s been around for a while, and we’ve always been aware of it.
When computing first became a commercial industry, women were actually at the forefront of it all. Programming was considered a woman’s job because, as The Guardian explains, “it was seen as deskilled and unimportant”. However, as technology progressed, people started to view the role of programmers rather differently.
“This quickly began to change as computers became indispensable in all areas of government and industry. Once it became clear that those who knew how to use them would have great power and influence, female programmers lost out despite having all the requisite skills.”
Just like that, the script was flipped. Women were no longer the frontrunners in tech, but the ones who couldn’t quite hack it in a male industry.
As recently as 2005, the then-president of Harvard University, Larry Summers, claimed that women made up a small percentage of those in STEM industries because of “innate” differences. In the same year, only 12.5% of tenure offers made by the university’s Faculty of Arts and Sciences went to women. .
Now, because such a male-dominated environment, it is difficult – even with constant social change and improvement – for women to break into the industry. Sadly, a great deal of this is to do with harassment. As Forbes reports: “While not steady enough a decline to be a trend, there is compelling evidence to show that in areas where IT dominates like startups, a huge number of women experience sexual harassment and many women in IT experience fears of rape and sexual harassment.”
All around the world, women make up a minority of tech employees. In the UK, only 17% of people working in technological roles are women. In the US, only 20%. Surprisingly, it is India who is leading in terms of female representation – but even they have only managed 35%.
Likewise, white employees have consistently been favoured over people of colour in tech.
As reported in CIO, a report from Open MIC found that “black, Latino and native Americans are unrepresented in the tech industry by 16 percent to 18 percent compared to their presence in the U.S. labor force overall.”
Furthermore, a survey cited from Indeed found that, of 1,002 tech professionals, 52% of non-white/non-asian employees had been made to feel uncomfortable by non-inclusive behaviour at their current place of work. When asked if they had ever felt uncomfortable because of their race, gender, age, religious affiliation or sexual orientation, 41% of non-asian/non-white professionals say they had. Meanwhile, only a quarter of white people could say the same.
And it wasn’t that long ago that racism was just a blatant fact in all workplaces. In 1994, The Bell Curve, which attempts to link race and IQ, was published by Charles Murray and Richard Herrnstein, subsequently sparking a controversy that is still being debated.
What this amounts to in tech is an industry predominantly controlled by white men. This is not to say that those white men are solely responsible for the bias in tech, or that they are not qualified for their roles. Not at all.
What it does mean, however, is that the technology they go on to create is informed by their knowledge and experience. When they program facial recognition software, they program it to recognise faces like theirs. When they teach it what a ‘typical’ employee profile was – their race, their gender, their educational background – they taught it the bias that they had been an involuntary part of.
With that in mind, then, is it any wonder that Amazon’s AI-driven recruiting tool consistently favoured men over women for tech roles? Or why Facebook’s advertisements showed racial discrimination when selecting targeted audiences? Or why several algorithms designed by top companies were discovered to have error rates that were 35% higher when identifying the gender of dark-skinned women when compared to light-skinned men.
The technology itself is not biased, but the long history behind it is.
How do we fix it?
Thankfully, there is a solution: hire more minorities.
Alright, the solution isn’t that simple – there are an innumerable amount of problems that need to be tackled before we can even begin to level the playing field in tech industries. The first is to dismantle the bias that exists in schools and towards young children.
A recent study by the Institution of Engineering and Technology found that parents’ beliefs about jobs being either male- or female-suited are discouraging girls from pursuing a future in the science, technology, engineering and maths (Stem) sector. These biases are compounded by teachers who – despite there being no evidence – think that boys are more proficient at certain subjects, and girls at others.
It is the responsibility of technology companies and authorities to let young girls and women know that there is room for them in the industry, and they are capable of succeeding in a tech role if they want to.
Then, within workplaces, stronger efforts must be made to hire smarter. Medium provides a helpful list on how to do this, and their points can be broken down into:
- Interview differently
- Provide interest and support groups within the company
- Curate a healthy culture of acceptance
- Respect everyone equally
- Engage with the wider community
By interviewing differently (ie basing job offers on proven skills assessed via unbiased tests and challenges), companies can essentially hire blind, or without bias. Then, once new employees are on board, businesses can better support those who are women and/or BAME by actively encouraging relevant interests, support, and respect.
More than anything, it is important to validate women and minorities in tech roles.
As Medium also notes, “Out of 557 female scientists interviewed by Harvard Business Review, two-thirds reported having to prove their skillset time and time again. And unfortunately, this constant need to prove their worth has driven many women out of the career.”
In advanced industries, workers from minority backgrounds constantly have to prove themselves. Being highly ranked at a top company still isn’t enough, apparently, and people will still find themselves being judged or challenged.
“Unconscious bias is leading to employees feeling insecure or personally attacked in the workplace,” Medium states. “An effective company ensures the well-being of all employees, regardless of demographic or circumstance.”
So, by normalising a diverse workforce and demonstrating that, actually, women and people of colour can and do achieve as much as their white male colleagues, companies can gradually stamp out this unfounded bias.
If absolutely nothing else, having biased technology simply isn’t useful. In fact, it could even be dangerous.
“Algorithms, technological devices and artificial intelligence, when badly designed, expose us to risks as much as an unhygienic meal or faulty pacemaker,” explains Lizzie O’Shea for Verso. “If used by government, they might even compromise administrative processes that govern our social security or judicial decisions about our personal liberty. We need to think about them as products that are designed and can be modified, and reject the arguments put forward by the tech elites to deny responsibility for their impact and blame users instead.”
The whole point of AI is that it is supposed to make our lives easier, and it can’t do that if it is unable to be neutral.
However, it is more than possible to reach that goal. With more women and minorities in the industry, and more lived experience to influence the technology that comes out of it, we won’t just have more accurate machines, we’ll have computers that are better overall.
It’s been consistently proven that diverse workplaces work better. They solve problems faster, make smarter decisions, and show an increased level of creativity. Think of what we’ve achieved so far with a relatively homogenous workplace; how much better would it be with the added strength of diversity?
The effort to boost broader representation in the tech industry can only be a good one, and is sure to improve on the incredible advances that have already been made.
For more technology news like this, insights into the world of Microsoft, and unique job opportunities, make sure to sign up to the Elliott Browne community.