In some of the circles discussions artificial intelligence (AI) and machine learning, there is the recognition it is not ‘neutral’. In fact, AI is often quite biased, repeating if not amplifying the bias in our society.
We see this in the USA regarding young, black men. Datasets from the police are used to help machines ‘learn’ and ‘analyse’ crime data. However because there has been historical discrimination towards young, black men, the datasets are heavily biased and racism amplified.
We see this when AI is used by HR departments of organisations to analyse applications or determine compensation packages for employees. Historical datasets are used to ‘train’ the AI, however historical datasets are biased against women in a multitude of ways. Discrimination and the gender pay gap is perpetuated.
Thankfully there is growing awareness of bias, sometimes unintentional, in technology and ‘going digital’. I’ve been struck recently by how I often only think about gender bias and bias against people living in an offline world or without a device. And yet, to my shame, I rarely think of discrimination towards people living with disabilities. And not just physical disabilities, but also mental and especially in children.
In many of the communities we work, disabilities are hidden as there is a sense of shame attached to the individual and their family. In conflict and post-conflict contexts, there is physical disabilities (e.g. loss of limbs) in ex-combatants (adults and older children) is slightly more ‘acceptable.’ But children born with disabilities or adults with mental disabilities are often hidden away.
We need to move a million miles beyond simply recording that a household includes someone living with disabilities when we register them. We need to move beyond always opting for the guardian model as guardians do not always exist. What does moving to digital identity mean for those living with mental disabilities eking out survival in a slum?
And yes, I can hear the question about don’t people in slums have bigger problems than digital identity. Of course they do, but if we are building a world in which everything works through digital systems then we are only amplifying inequality if we only think about those in the suburbs with smartphones.
As the work of civil society organisations continues to ‘go digital’ how are we not perpetuating and amplifying the discrimination already faced by people living with disabilities? How can we use the process of ‘going digital’ upend the discrimination and help us as societies to be inclusive?