Defining fairness in the tech industry

The discussions around digital ethics cover a vast range of themes, from the functionality of the tech itself to the business practices which govern them. But one word which keeps recurring in almost all these conversations is fairness. 

For this to be meaningful, we need to know what is meant by fairness and how it applies to technology. 

To get etymological, fairness is officially defined as treating people equally without favouritism or discrimination or without cheating or trying to achieve an unjust advantage. 

Using this definition, the tech industry cannot be said to be entirely fair. This may appear a big claim but there are a couple of reasons I reach that conclusion – including, although not limited to: 

• Referral networks. While some start-ups make hiring decisions based at least partly on suitability for the role, there is no denying favouritism plays a role as well. People with contacts in an organisation have a much higher chance of becoming the favoured candidate. 

• Representation. Too many companies do not represent the breadth of the diversity of their customers or even the cities they operate in. It is hard to argue that phenomenon is anomalous given it occurs systematically across developed markets and across industry. 

But what are the costs of this lack of fairness and how does it impact businesses?  To start, where there is discrimination, it is more likely policy intervention will be needed through regulation, therefore increasing companies' compliance costs. Other costs to companies include retrofitting bad code which is a lot harder and usually more expensive than doing it right in the first place. Beyond that how does one calculate the brand damage associated with discrimination? More broadly, we just get worse products and fail to act upon the evolving economic potential of marginalised groups. This has real tangible consequences for businesses who risk producing products, services and solutions that do not meet customer expectations, or scale effectively. Ultimately, when diverse voices are excluded from the decisions that shape these views, we push divisiveness to the forefront of our lives. 

So where do we go from here? Progress is being made and more is being done to create more inclusive environments within the tech sector and its constituent components. Still, some parts of the industry are more progressive and move at a faster speed than others. 

In equitable engineering, which is ensuring that technology does not discriminate against certain groups, we see increasing efforts to adapt to users with disabilities and the adoption of small measures such as alt text. This focus on assistive and accessible aspects of digital means that we can include more people in the conversation. 

When it comes to race and ethnicity, fairness is also not only about hiring and representation in the industry, but about data and algorithms. As we advance in machine learning and data science, we are ever more reliant on data which can often be representative of an unjust and inequitable reality through which it was obtained, collected or generated. Efforts are being made by scale-ups and big tech organisations to bring attention to this, mitigate against it and include new voices into the conversation when it comes to what needs to be done about it. Nonetheless, these are just the start and we need to make greater progress.  

Representation is just the proxy, not the end goal. Having a more diverse team or business results in better decision-making, reduces risk and spurs innovation, but simply having folks sat at the table is not enough. It has to be accompanied by the idea that those folks are empowered to be involved, dissent, challenge and contribute. It’s in belonging, equity and inclusion where the hard work remains, should the tech industry really be fair.