What is Data Ethics?

Data Ethics can be described as a code of behaviour, such as what is right and what is wrong. Such as Data Handling, Algorithms and Corresponding Practices. See this article by Dataversity for more detail.

There is much to be said about Data usage and AI ethics, especially since the Cambridge Analytica scandal. However, we often forget that it was mathematicians at the helm of 2008 financial crisis and the nerds at Google that trained the infamous “racist” algorithm.   

The above experience became viral on Twitter.

The above seems like a harmless recommendation, right? Well, imagine a scenario in which self-driving cars fail to recognise people of colour as people and are hence more likely to hit them! Now do you get my point!?  
Algorithmic products are based on Machine Learning or more correctly put, statistical models which rely on the input of Data. If the data sample is biased or the math is not properly applied, we end up with one of the worst irritations known to man, inaccurate Netflix recommendations! 

Recently we all watched the Netflix Documentary “The Social Dilemma”, and if you haven’t, stop right there and go and watch it.  The Social Dilemma, even though it somewhat sensationalised the threat of Big Tech, did get one thing correct that our Data is worth a lot and if we are not paying for a service or product then we are the product being sold to advertisers.  

All these examples are only for the Silicon Valley giants or Governments, right? Wrong, SME can also fall into these pitfalls leading to GDPR transgressions, or inaccurate usage of statistical modelling relying on fuzzy data. 
Having a good strategy of Data governance and ethical use policy will only support an organisations ethical obligation, develop consumer confidence and support more accurate use of data.  

Our recommendations to get started, beyond developing a policy that will be saved on your servers somewhere, are:  

  1. Be transparent: Transparency on the data you collect on consumers is a legal obligation so having the correct processes is crucial.  
  1. Audit your data assets: Audit what data you have and assess how you could collect better data or make better use of the data you have.  
  1. Ask questions:  Always criticise the use of data and the use of mathematical models; for example, are the samples your Data Analysts are using accurate enough or how much bias does the training data for a model have?   
  1. Be a leader: as many data challenges are new don’t be afraid to investigate and tackle them to get ahead of the competition.  
  1. Train Staff: Keep staff trained and up to date on policies and the latest use of practices in data governance.  

As you develop your processes and understand what your business requires an excellent resource is The Open Data Institute, however, a lot of their focus is on Ethics and governance. For better products you will have to rely on developing talent.  

So, ask yourself “How can I develop talent?”. Get in touch with our team today and let’s talk about building data talent throughout your team. 

Rehmi 
Head of Data Analytics
Imran.rehmi@gbs-ltd.co.uk  
01246 925923

Stephen Archer 
Business Development Manager 
stephen.archer@gbs-ltd.co.uk 
01246 925923 

Click below for more information on our Data Analytics Apprenticeship Programme.