How can tech help to promote diversity in business? We looked at the various types of biases and the importance of diversity in the workplace.
Have you ever stopped to think about diversity in the business you work for? Consider gender diversity, ethnic diversity, and racial diversity. Are different groups given representation and better still, equality? What is the average profile of an employee, of managers? Can you see all walks of life reflected in your employees, in your leadership?
There are many benefits of diversity, but it’s less likely to be found in a company than you might expect. Diversity and inclusion is certainly a hot topic at the moment, so why is equal representation still proving so difficult to achieve? We wanted to look into the business case for diversity and tried to uncover why the workplace still isn’t becoming more diverse.
What is diversity in the workplace?
In simple terms, diversity in business means hiring a varied and diverse workforce. This covers people of different ages, genders, social, educational, racial, and ethnic backgrounds.
In reality, true diversity in the workplace runs much deeper than this. Companies should be diverse and inclusive, creating teams of varied individuals who have equal opportunities. But the truth is, the business world is biased and no matter how much technology you integrate into your business to overcome bias, ultimately major decisions are handled by human beings. Emotional, and frequently irrational, human beings. Wherever humans are involved in the business processes, there’s always scope for bias to creep in and deviate from rational decision making.
So, is it possible to create true diversity in the workplace and reduce unconscious bias whilst we live in a biased world? And is it really technology that’s going to get us where we need to be?
Types of Bias
There are many types of bias, but they can usually be grouped into Emotional and Cognitive:
We make decisions based on our emotions, instead of logic. Emotional bias examples in business include:
- Bandwagon effect: We adopt ideas because others are, rather than analysing the information carefully. Dangerously commonplace in team meetings.
- Empathy gap: Thishappens when people in one state of mind fail to understand people in another state of mind. E.g. if you’re offended by something and you don’t understand why others aren’t.
- Hyperbolic discounting happens when people make decisions for a smaller reward sooner, rather than a greater reward later.
- Negativity bias is the tendency to put more emphasis on negative experiences rather than positive ones.
- Overconfidence is when some of us are too confident about our abilities, and this causes us to take greater risks in our daily lives.
Where our biases result from incomplete information, or our inability to analyse the information available. Cognitive bias examples in business include:
- Confirmation bias: The classic one. We only listen to information that confirms our original theory, discounting new, conflicting, information.
- Conservatism bias: Where pre-existing data is deemed more valuable than new data.
- Anchoring bias: Using the first piece of information we hear usually affects our decision making.
- Recency bias: Future decisions are based on memorable, recent events, even if they are one-off events, because they are easier to remember.
- Frequency illusion: This occurs when a word, name, or thing you just learned about suddenly appears everywhere.
Any prejudice can have the potential to affect your business operations. Whether it involves future-planning, the day-to-day, or even the hiring process, nothing is immune. Whilst we assume a lot of these biases are subconscious, other far more serious biases – such as racism and sexism – are conscious, illegal, and widespread.
What are the 4 types of diversity?
The four types of diversity are internal, external, organisational, and world view. Exploring these will help us to understand what true diversity in business looks like.
- Internal diversity types are born with an individual and are very difficult to change. For example, ethnicity, age, race, and gender.
- External diversity types refer to characteristics that a person is not born with, that can be influenced, changed, or developed by an individual. These may include appearance, education, and interests amongst other things.
- Organisational diversity refers to factors belonging to any particular workplace. A few examples are work location, job function, and union affiliation.
- World view covers the way life experiences affect an individual’s view of the world. Diverse world views may be political, historical, or based on cultural events.
Now we know what it means to have a truly diverse workforce, do we even know when we’re being biased? These stats illustrate the current state of diversity management and inclusion:
- Over 1/3 LGBT+ staff have hidden or disguised their sexuality at work because they were afraid of discrimination.
- There are only five black CEOs on the Fortune 500 2020 list —just 1%!
- According to the CIPD, 51% of HR professionals in the UK were found to be biased against overweight women – and were unaware that this was the case.
- Only 7% of the world’s languages are reflected in published online material, with English dominating at 50%.
This certainly highlights the business case for diversity and inclusion. Gender diversity, racial, and ethnic diversity are far from where they should be. But why is diversity important in business?
What is the importance of diversity in the workplace?
There are many benefits of diversity in business. When a company is truly diverse and inclusive, both employees and the business as a whole can benefit. Companies that employ more varied people automatically get the benefit of different perspectives. This increases creativity and productivity. Employee satisfaction increases and people perform better when employees feel included as individuals.
The case for diversity doesn’t end there. Recent research conducted by HRWins found that nearly 60% of U.S. and U.K. employers surveyed believe there is a direct financial impact on the business when an organisation is competitive in inclusion and diversity. So true diversity is likely to have financial returns too.
Why the solution lies in the people who wield technology
The recently released Netflix’s documentary Coded Bias highlighted the many ways in which, while AI technologies are incredible tools that have improved our lives, they are as prone to bias as the people who create them. This is why a more inclusive workplace needs to begin at a human level, not simply with the implementation of new technologies. Shalini Kantayya, the film’s director, told New Scientist “I hope [Coded Bias] will be a kind of Inconvenient Truth of algorithmic justice, a film that explains the science and ethics around an issue of critical importance to the future of humanity.”
So how can an algorithm have a bias? Because any algorithm has a human behind it who can unknowingly project their own unconscious bias. The National Institute of Standards and Technology (NIST) conducted research across 189 algorithms developed by companies including Intel, Microsoft and Toshiba.
The purpose of the research was to see how accurate facial recognition software was on average, and if there were common failings. The results were concerning, to say the least.
NIST’s research found that across the board all of the facial recognition software they tested was far less accurate when identifying African-American faces and Asian faces compared to Caucasian faces.
How did researchers test the algorithmic bias?
False Positives and Negatives
One way in which NIST tested the bias of algorithms was in one-to-one matching. This is where the facial recognition software tries to match photos to identify if it is the same person in both images.
For example, one-to-one matching can be used in criminal proceedings to match a suspect recorded on camera with photographs of suspects.
Many of the algorithms tested falsely identified Asian and African-American faces 10 to 100 times more than Caucasian faces. African-American women were the most likely to be misidentified, overall.
Congressman Bennie Thompson, chairman of the US House Committee on Homeland Security said “the administration must reassess its plans for facial recognition technology in light of these shocking results?”
Why is this important?
Facial recognition is increasingly common in our everyday lives. Many of us use facial recognition to log into our phones and computers, and airports have even begun to install facial recognition software so that you can guide yourself through customs.
As with any new technology, some flaws can be expected. For some time, facial recognition software has struggled with sunglasses and reading glasses. This is because the reflection from your glasses prevents the software from being able to get as many data points as it needs to build an accurate portrayal of your face. This might be an inconvenience but there are other ways to access your phone and researchers have continued to work on the problem.
NIST’s research is more troubling because it highlights a systematic problem with facial recognition and race in software that is used in law enforcement.
While some US states have banned facial recognition because it isn’t reliable enough, it is still being used in other states. Amazon, who sell the facial recognition software Rekognition to US police forces, declined to take part in NIST’s study. Previous research from the Massachusetts Institute of Technology had already reported that Amazon’s Rekognition struggled to identify women with darker skin, although Amazon denied the claim and called it misleading.
This is an issue for the police force, who can’t work accurately because of non-inclusive technology, and the people it will impact.
But why does technology still struggle with bias?
According to The Algorithmic Justice League, much of the problem lies in a lack of inclusivity at tech companies themselves. They report that “in the U.S, the teams designing all systems are not inclusive. Less than 20% of people in technology are women and less than 2% are people of colour.”
While there is no doubt that developers want to create cutting-edge technology that can best serve everyone, a lack of inclusivity within teams means there are only very specific viewpoints and life experiences being catered to.
How does this lack of inclusivity in teams manifest?
The problem with a lack of inclusivity in technology isn’t just limited to facial recognition either.
Motion detection soap dispensers have been found to be unable to recognise darker skin tones. This was of particular concern for an American-American guest of a sci-fi and fantasy convention. Fitzpatrick reported “I tried all the soap dispensers in that restroom, there were maybe 10, and none of them worked. Any time I went into that restroom, I had to have my friend get the soap for me.” This was a frustrating experience for Fitzpatrick, but imagine if he had been working in hospitality where the simple ability to wash your hand is essential and technology prevented him from being able to easily do so?
And voice recognition has repeatedly had a problem with understanding certain accents. One user reported that Alexa couldn’t understand her mother-in-law’s Nigerian accent. In fact, Alexa wouldn’t respond to any commands until her mother-in-law put on a stereotypical British accent. Even then, Alexa struggles with many regional English-speaking accents, specifically a regional Liverpool accent and in the wider UK, Alexa especially struggles with the Glasgow and Cardiff accent. Of the British accents Alexa does do incredibly well with, many are Southern cities where much of the UK’s technology companies reside.
How does this technology bias impact the workplace?
Many of the examples we’ve discussed here discuss how technology can impact us on our personal level and for many, it may seem a case of simply waiting for the technology to improve.
But, as many technologies become increasingly commonplace, they become necessary tools in the office. That means until the technology improves, it will hold back certain people unfairly. As shown in our infographic, Zoom is known for erasing black faces when it can’t find a pale face in the room. Try having an online business meeting when you can’t even get your face to show up on a video call.
Twitter’s algorithm has also been found to default to white faces over non-white faces when cropping images. This makes it more difficult to be seen online if you don’t have the right skin colour. The ability to be found online is increasingly important in many industries and this is just one way social media makes it harder for some people to be recognised.
Companies will also find themselves left behind if they cannot develop inclusive, accessible content as they’re limiting their own customer base. Companies that value diversity are proven to make more money.
How to promote diversity in the workplace
The benefits of diversity are clear, so how can we promote diversity in business?
- Bias blindspots: Seeing bias in others is easier to spot than in yourself. Make sure someone is assessing your bias regularly – and take regular tests. The Harvard Implicit Association Test is worth a try.
- Name-blind recruitment: This is already in place in the UK Civil Service, HSBC and the BBC.
- Better training of algorithms: In order to create less biased tech, we need to create algorithms that are trained with a full spectrum inclusion. That is, a more diverse data set.
- Get out of your comfort zone: Develop a working relationship with a counterpart who comes from a different background – it can only be a good thing for you and your business.
- Be more visible to your staff: Make your HR data more visible and easy to analyse. Not just for HR leaders or top executives but also for line managers who are increasingly being held accountable for performance.
- Keep improving yourself: There are a multitude of apps that can offer guidance on language and tone of voice, even providing “inclusion analytics” in real-time.
- Slow it down, smarten up: Use smart, or “slow thinking” to make decisions instead of making judgements on little to no data.
How technology has made us more accessible
While we’ve focused on some of the dangers of technology here, it’s important to note the many ways that technology has already made the workplace more accessible. Much software has been designed to make computers accessible for those with visual impairments, making offices more navigable, user-friendly spaces.
It’s about continuing this momentum and learning about the biases we may not be aware of that impact how accessible our workplace is to others.
It feels like catch-22 but it has to start with being better humans – knowing, owning, and correcting our collective biases. In 2018 alone, more than $4 billion was invested in the HR technology sector, yet relying on technology alone will not solve the issue of bias in the workplace.
Diversity is no longer an issue of compliance, it’s imperative to a modern, successful business. Open your channels of communication from the very top so your team learns how to give and receive feedback. Remember that real diversity and inclusion is about fostering an environment where all employees feel heard and appreciated.