ith data being the most valuable asset within the market today, the manner in which companies are collecting data have been pushed so far that they are now affecting people’s mental health.
But with all the money and power that comes with it, who wouldn’t want to capitalise on the data game opportunity.
Technology in today’s day and age consists of a wide range of automated processes in order to reduce the need for manual human labour and improve overall efficiency, machine learning and AI being the leading force of that. The concept of machine learning allows a system to use historical statistical data in order to work out the potential of an outcome in the future. Whereas artificial intelligence refers to the development of computer systems able to mimic human behaviour via things such as visual perception, speech recognition and decision-making.
Social media uses these machine learning algorithms based on historical user behaviour. These algorithms powered by supercomputers then detect and analyse the behavioural patterns of users to keep them on the desired social media application. The Social Dilemma being the main advocate for exposing what social media and social networking companies do in order to manipulate the human brain and how these algorithms have taken it further. Ultimately, questioning its greater impact on society and its moral and ethical stance.
To answer this particular conundrum, we have conducted an interview with Mark Wilson who has 23 years of experience in the data industry. Mark started out in software development and then moved into the backend of data. He has been a business intelligence consultant for over a decade and has developed two start-ups. One in the digital marketing sphere, the second in data science, AI and in data visualisation. Mark currently works as the head of data and analytics at a super application called Ayoba.
Ayoba is a local All-in-one South African application which allows users to text, read channels, listen to music, view sports, fashion as well as play games. Because this application contains all these various sub-applications, there is a lot of data that can be extracted and analysed.
How is your data collected? At Ayoba, there are two main ways of data collection. The first channel is the collection of Event information on the actual app. This is monitoring where a user clicks and taps and how long a user spends looking at an article, listening to music orchatting to their friends. This data is usually compiled, in apple and androids, using google analytics. The second channel is via MMPs. MMP, which stands for Medium Measurement Partners, are used to understand the attribution of instalments.
This means that they track where a user engages to install the app, whether it's via a link on the Ayoba website or an Instagram influencer post and there after the net downloads via those individual avenues. MMPs are usually used to decide where to spend money for marketing and with all the collected data, these companies can start to build a digital persona tailored specifically to keep you engaged.
A common use for data information like this is for a technique called Mood Targeting. The technique combines certain behaviours and triggers in order to ascertain the mental state of the user. This is done in order to market appropriate ads or elicit certain triggers to keep the user engaged. Mark stated that when analysing data they look at user behaviours such as the time of day that someone streams music and the type of music that they stream, as this might give a glance into the user's mental state. This allows them to advertise something that targets the user's current mood.
PHOTO: COURTESY OF APPLE
Although Ayoba does not have a mood targeting feature encoded into their app, they lure users to use more features of the application and by doing that retain users attention for extended periods of time. Being a SuperApp which has multiple avenues for users to use and visit, they are then able to pinpoint which channels of the app the users enjoy most. This means that if the user is browsing through their channels and decides to read a recipe, the application would automatically suggest listening to a playlist they might like while the user cooks.
Machine learning is an integral part of what makes persuasive technology. Persuasive technology is the computer system design that gets people to use as many of the aspects of an app as possible. Subsequently changing people’s behaviours and attributes. This technology follows a simple formula, B=MAT (Behaviour = Motivation, Ability andTriggers). These 3 factors need to be met before a person changes their behaviour. An individual requires sufficient motivation to perform an action, then have the ability to perform the action easily and ultimately trigger to perform the set action. At Ayoba their form of Persuasive technology is called LCMs, Life Cycle Messaging. LCMs are notifications which pop up, stating the user hasn’t used the app in a while or that they released a new feature.
Ayoba’s team questions, “what do this cohort of users do when they follow a recipe?” For context, Ayoba’s knows the user likes listening to music based on their collected behavioural data. They will then send a push notification asking if they want to listen to the music while they cook. Their love of music will be their motivation, the ability to listen will be the easily clickable push notification and the trigger is the actual appearance of the push notification. Mark states,
“
They [the user] do what you suggest them to do[the behaviour]. The question of what they would have done outside of that suggestion [the trigger] is the behaviour that you want to change, which is gone because the training dataset has changed.”
The only times technologies such as these become morally questionable are in applications such as TikTok. With its card-based interface and very brief cycles of attention, it targets human beings’ ingrained behaviour to seek outcomes for something being watched. TikTok short circuits this process by giving the user a payoff for 15 seconds of attention.This payoff is a hit of dopamine. Dopamine is a neurotransmitter in the brain that acts as a reward signal and contributes to reinforcement, motivation and learning. Without putting in any effort and sitting and looking at something for 15 seconds, one gets the reward of a dopamine hit which the user can do repeatedly.
This creates an addiction in order to keep users on the application. The way this keeps users on the application is that the dopamine hit becomes smaller and smaller with each scroll which subconsciously makes users fight harder and harder for it. “This makes the application sticky and stickiness is defined as: To what degree is this application part of your daily routine?” says Mark. For Ayoba, they look at which individuals open the application more than 10 times a day and then what is the first thing they do on the application. They can then use this data to bring back more people to their application, more frequently.
“
“No one cares what they publish as long as it results in shareholder value,”
Thus, horrific scandals such as Cambridge Analytica took place. By using machine learning algorithms, personal data and audience segmentation techniques from up to 87 million Facebook users, they psychologically targeted certain groups of people following the American election in order to influence their behaviour, emotions, and attitudes. “Being divisive is more profitable, just like how negativity is more profitable than positivity,” Mark concludes.
Not many ethical concerns are addressed when it comes to the manipulation of the user, the goal is to beat the other competition.
All in all, one cannot say that these technologies are either good or bad for they can be used in both positive and negative ways. Although social media may be using harmful and divisive engagement techniques which can result in negatively affecting the mental health of many individuals, it is the user who chooses whether to engage or participate in online applications and social media and how they go about accomplishing that.