Big Data for Big Impact – but not only a positive one

Technology has changed and keeps dramatically changing our everyday life by transforming the human species to advanced networked societies. To celebrate this digital revolution, 17 May is dedicated to the “World Telecommunication and Information Society Day” (WTISD-17).

The theme for this year’s celebration is “Big Data for Big Impact”. Not so surprisingly, the buzzword “big data” echoes in our daily commutes over the internet world. The chosen theme focuses on harnessing the power of big data to turn complex and imperfect pieces of data into a meaningful and actionable source of information for the social good.

Big data has a potential to improve society – much like electricity or antibiotics. From health care and education to urban planning and protecting the environment, the applications of big data are remarkable. However, big data comes with big negative impacts. Big data can be used – by both advertisers and government agencies – to violate privacy. The power of big data can be exploited to monitor every single detail of people’s activities globally.

With 29 million streaming customers, Netflix is one of the largest providers of commercial media in the world. It has also become a trove of data for advertisers as it collects data on users’ activities – what, when and where they are watching, what device they are using, when they fast-forward, pause or stop. Just imagine a representative of Netflix sitting behind your couch, looking over your shoulder and making notes whenever you turn on the service. This applies to many online services, such as Google, Amazon, Facebook or YouTube.

Mass surveillance initiatives by intelligence agencies such as the US National Security Agency (NSA) and the UK Government Communications Headquarters (GCHQ) take this power to the next level to knock down every bit of personal space. Without big data, the scale at which such profiling is done today would not be possible.

It is very tempting to use the benefits of big data for all sorts of purposes. Hiring new employees based on their social media activities, granting insurances based on fitness tracker data, airport security check ups and future crime predictions based on cell phone call logs, to mention a few. But there are some fundamental problems with applying big data to services.

The first problem is that, knowingly or unknowingly, we all have biases when making decisions. If decisions made by millions of employers, policemen or judges over a long period are collected together, it brings in all those biases, on a bigger scale. Big data may just refer to a large chunk of unstructured data, but the insights deduced from it will rely on machine learning – which accumulates all possible biases, such as gender and race. Algorithmic decision-making could turn out to be more biased than ever before, which would have a terrible effect on the society.

The second problem is the error rates: A study on automatic face recognition software found that the error rates can vary between 3% and 20%. This means that your face could match with one in the database of potential terrorist the next time you go to the airport and you could be pulled out for questioning or get into even more trouble. This is happening in the international airport transit on a daily basis. It is not possible to create 100% accurate models, and every time the assumptions are made on a missing data sample, the errors are inevitable.

Therefore, when dealing with big data, it is crucial to be extremely cautious about the quality and sources of the data, as well as about who can access it, and to what extent. If a data set stemming from diverse sources is handled with special care and anonymised thoroughly to protect privacy rights, big data can be used to solve complex societal problems. But if it is left unregulated or not properly regulated, and not tested for its fairness and biases, it can pose a serious threats to our human rights and fundamental freedoms.

EDRi has fought for the EU General Data Protection Regulation (GDPR) to regulate this practice. Now EU Member States are implementing the GDPR, and it is up to them not to abuse the weak points of the Regulation to undermine the protection of the European citizens’ data.

Video by EDRi member Privacy International: Big Data

(Contribution by Siddharth Rao, Ford-Mozilla Open Web Fellow, EDRi)

Originally published at EDRi.org