Privacy in the Age of Data Technologies

Privacy is a fundamental human right. However,  for many, the words to articulate exactly what it is can become elusive, especially when it comes to data technologies such as Artificial Intelligence (AI). These technologies cause us to  reflect about what privacy really entails, especially because of the fast pace with which data technologies are evolving.

We particularly like the definition by Bennett & Raab of privacy as a boundary between individuals, individuals & their governments, and individuals & corporations. This boundary shifts depending on context (e.g., culture & the application). It is subject to negotiation, such as for public good reasons in the context of  security and safety. A good example is security searches whenever we go through airports. 

Data technologies have significantly blurred these boundaries, although we may not always be privy to the exact ways in which this is occurring. In exchange for services such as navigation, convenience of shopping from the  comfort of our couches at home, or ease of finding movies and music to stream that are to our particular tastes, we inadvertently also give away a lot of data about ourselves that can (and often are) used against us. Strangers can now tell when we are not home through our social media posts, leading to increased potential for home burglaries when we are not home. From a prediction or inference perspective, the ability to predict our behavior further demonstrates the potential for privacy violations. For example, Target (the supermarket chain) was able to accurately predict a teenage girl's pregnancy before she knew about it and was able to tell her family. This is one of many examples of why data mining has received a bad reputation. And the purpose of data collection is the same for AI: to uncover/learn (hidden) patterns in the data. Consequently, privacy risks remain the same, if not greater.

AI, in particular, is very good at inferences at scale. Way back in 1889, in the book “The Sign of Four” written by Sir A.C. Doyle, the character, Sherlock Holmes, says “While the individual man is an insoluble puzzle, in the aggregate he becomes a mathematical certainty. You can for example, never foretell what any man will be up to, but you can say with precision what an average number will be up to. Individuals vary, but percentages remain constant.”  AI operates through the very same principles and, therefore, this quote resonates even more so today because of the massive amounts of data collected  and augmented by computational power which has, itself, has been on an exponential growth.

The public, however, is catching on and becoming more aware of their data rights. This is evident from enactment of regional data protection laws and AI laws such as the EU’s GDPR, AI Act, and Digital Services Act. Africa is also not left behind with 45 of 54 countries in Africa having data protection laws in place. 4 of the remaining 9 have draft bills in progress as well. Laws, however, are only one of three dimensions of privacy, the other two being technological, and cultural. All three intersect and are invaluable in balancing the benefits and privacy risks of AI (and data technologies in general).

On the technological front, there are various ways to preserve privacy in data technologies. This can range from rudimentary techniques such as data minimisation,  anonymisation (e.g., k-anonymity), and restricted access, to quite sophisticated methods such as differential privacy and homomorphic encryption (see figure below). On the cultural front, privacy becomes a bit more abstract when one has to understand how individuals in a community prefer to relate to each other, as well as to their government and various corporations operating in a specific locality. This is also where debates such as individual privacy vs group privacy, or Ubuntu formulations of privacy can take centre stage.

At EthicEdge we see opportunities for privacy preserving data technologies that also reflect cultural values in the African context. We are making it our mission to help organisations, governmental institutions, and civil societies engrain privacy into their use and development of data technologies. We are co-designers and co-thinkers, rather than sideline critics.