By Prof. Bitange Ndemo
Uttered by US civil rights activist Martin Luther King Junior over 50 years ago, these words remain relevant today as new technologies take the centre stage in Africa. Rev King Jr. was fighting for the rights that protect personal freedoms from infringement by governments, social organizations, and private individuals.
One of the risks associated with emerging technologies is the possibility that human interaction and the algorithms that drive them will hardcode bias into their decisions. For example, some codes can systemize inequality. Ironically, it is the ability of these technologies that can reduce human decision-making biases and help fight technology-driven discrimination.
Such conveniences are the ones that give us the false comfort. We forget that we are in challenging times where Africa is not readying itself to face the test of the emerging fourth industrial revolution. At a conference in Nigeria last week, one of the African diaspora experts narrated how she attended an Artificial Intelligence (AI) conference in Europe and as the organisers were showing them new facial recognition security solutions that that can also estimate the age of the subject, they had each guest test the system’s accuracy in turns.
When it was her turn, the system did not respond. It had not been trained to scan black people. Fortunately for her, this was just a demo. In some airports, systems have been known to profile and pick specific ethnicities for detailed searches.
In another case that happened in the United States, the hand drier in washrooms could not respond to people with dark skins. Here too the algorithm did not have such provisions and hence the discrimination. The examples seem to be trivial but they cause unnecessary anxiety especially when observers assume that the user is ignorant and therefore unable to the use a simple hand drier.
These stories will continue to pile up until Africa becomes a major force in technology. That is when developers of new solutions will not forget the existence of some 1.3 billion living people of African descent.
There are many issues that we don’t understand if there are inbuilt biases. It is only recently that the debate on race-based pharmaceuticals emerged in the US. Duke University geneticist and bioethicist Charmaine Royal in his article “Will precision medicine move us beyond Race,” published on May 25, 2016 in the New England Journal of Medicine, acknowledges the genetic diversity that exist within racial groups and the similarities between different groups.
But when the first race-based drug, BiDil, was approved by the FDA to treat African Americans with heart failure, advocates without any data heralded it as a way to narrow health disparities between whites and blacks by targeting the group that suffered the most from the disease. Eventually, the drug was found to be ineffective.
The key takeaway from Royal’s article was that technological advances in DNA sequencing and analysing large datasets will continue to generate insights about the genetics underlying differences in drug response. The data deluge will only further highlight the pitfalls of using imprecise race categories to prescribe drugs.
The article concluded that moving beyond race-based drug prescriptions will depend on the ability to equip health care providers with the resources and training they will need to collect and make sense of more types of data.
That precision medicine is premised on the idea of improving health outcomes by generating and using many sources of personal data to more accurately group and treat patients. Royal’s parting words were “If the major challenges can be overcome, precision medicine could lead the way in reducing and ultimately eliminating the use of crude racial and ethnic census categories in drug prescribing.”
What we need in all these cases is to develop massive capacity around the emerging technologies, release data in formats that can be analyzed and collaborate with the rest of the of the world in solving Africa’s problems. That is how to unleash the power of data analytics.
The writer is an associate professor at University of Nairobi’s School of Business.