We must champion ethical data usage to protect society from exploitation

The world’s communication and by extension entertainment platforms are undergoing transformation. From the paper era to the information technology and presently, the mobile revolution. This global shift is exposing us to digital media’s pros and cons.

Depending on your income, social standing and job, your dependency on this technology varies, but generally, may affect several facets of your life.

Increasingly, the duration spent online on a per capita basis across Africa is rising, and is expected to widen as generation Z grows. The mobile phone, previously a preclude of adults, is now in the hands of teens. Percentage of young people with access to mobile phones is much higher than in the last decade and will likely treble by 2030.

While digital platforms serve useful functions such as informing, entertainment and e-commerce, concern is mounting over hidden monetisation agendas from insights gleaned during their utilisation. Targeted advertising in particular, has been singled out.

As more and more psychologists, sociologists and behavioural neuroscientists join digital and social media marketing units, the degree of sophisticated marketing means we stand no chance from the coming Artificial Intelligence AI onslaught.

One of the major questions being exploited is, “How do humans form decisions when presented with a raft of options?”

The ethical concern is whether technology has a right to influence our choices through behavioural manipulation. The basic neural networks exploited, essentially hinge around reward and motivation, as well as attention fixation circuitries.

Machine learning, a vogue topic currently, is a precursor of this AI future, whose basics are premised on engagement to learn, captivation to inculcate repeat visits, and higher consumption frequency as end points.

However, to make these choice-influencing moves, AI must first have solid building blocks. The most common one is equivalence classes formation. It’s very basic explanation is the linking of observations to make a pattern and conclusions.

For instance, if every time Gor Mahia plays, I go online (A) and within the same time span, I make a bet that Gor will beat AFC(B), an association deduced is that I am a Gor Mahia fan and a gambler. Repeat analysis of my digital footprint could inform more. The second level would then be behaviour reinforcements. Here, “nudges to consume” could be free punts every time I read Gor Mahia news. Tokens, “easy wins” (C) will then cement this class through the reward circuitry, creating a conditioned stimulus.

The complexity of the equivalence classes can span a couple of items, forming cross-functional bridges. Multiple platforms are typically necessary to make such, depending on the depth of information available to “the machines”. A concern on possible abuse of data to capture wider equivalence class-linkage bridges by tech giants owning multiple platforms arises.

We must champion ethical data usage to protect society from exploitation.