Category Archives: Security

COLLECTIVE COGNITIVE BIAS IN THE SOCIAL ERA

by Cristiano Galli

 

This article was drafted before the COVID-19 global issue exploded. Most recent events offer the opportunity address the subject of this article from a different perspective. Initially the title was “Are social media improving or reducing collective intelligence?”, but the globalized reaction to COVID-19 has triggered a new approach.

 

The level of connectedness of the globalized world is generating a new form of collective intelligence. Individual brains, like never before, through social media and information sharing in general, are linked and joined in the process of sense making, influencing social behaviors.

 

The ability of a social group to process information in a smart way has been defined by James Surowiecki as “Crowd Wisdom”. In his extraordinary 2014 book “The Wisdom of Crowds” Surowiecki explains the principles that enable a non-expert group of people to take wise decisions even in seemingly expert related context.

 

The question arising in this article is: is the globalized social information processing making the world smarter or dumber?

 

Let’s briefly introduce the principles mentioned by Surowiecki. He argues that a crowd requires four elements in order to be able smartly solve cognitive, coordination and cooperation problems. First: information has to be shared to the processing agents (in this case processing agents are people), without filtering. Second: processing agents need to maintain a balanced level of diversity and, third, autonomy in their information processing function. Fourth: there should be some form of aggregating function collecting the distributed processing into a collective output.

Lacking one or more of these elements, crowd decision-making process becomes less and less smart.

 

The first issue to address, which usually encounters expert resistance, is the fact that a group of non expert decisions could result smarter than the decision taken by an individual holding subject knowledge and expertise. For, as weird as it could sound at first glance it has already been scientifically proven in many fields: from guessing the solution of a cognitive problem to fixing the price of a product in a free market, and to regulate tax paying and traffic.

 

How can this phenomenon be explained? Crowd elaborated solution contains two parts: the solution itself and a related error. While effective solutions aggregate, errors – coming from autonomous and differentiated processing – elide each other, reducing dramatically the overall error.

 

Let’s try now to analyze this process from social media processing perspective. Information sharing happens to the speed of light, but it is not free from embedded biases. Information is hardly ever shared in the form of crude data, mostly is attached to already packed opinions that hinder the first principle for collective smartness. Shared information already contains part of the elaborated solution and so influence negatively the ability of the distributed agents to perform their processing function. Crowd processing becomes more like a resonance box for already processed solutions.

 

Another issue is the motive behind information distribution. Instead of distributing data, most modern media are entangled in a form of information sharing speed race. Speed is hampering the quality of the shared information. Recent studies on the diffusion speed and reach of fake news have demonstrated that fake news act on individual brain processing gaps, acquiring a much higher ability to be diffused in the globalized information system. Our brains are more prone to react to information that is perceived as a “different” and “strange” than information that matches more with the mainstream perception. Daniel Kahneman, in his Noble Price scientific production has distinguished two systems in the neural processing for decision making. System 1 is mostly connected to our most ancestral limbic system. System 1 is the automatic, associative, quick and energy efficient part of the brain. It has evolved in order to provide us with a quick decision making tool to survive. Think about one of our ancestors walking in the African savannah thousands of years ago. He saw some grass moving at distance. It could be a threatening animal or simply the wind. He needed a brain system to react quickly in order to maximize his ability to survive. System one kicked in, addressing the grass movement as a threat and supporting his evolutionary behavior to escape. He might be right (it was a saber tooth tiger) or he might be wrong (it was just the wind), but in each case, it is survival had been guaranteed. The more advanced System 2 is instead the rational and cognitively more reliable decision making tool. System 2 is less prone to mistakes, very effective in complex decisions, but it is slower than System 1 and consumes a lot of energy. Fake news diffusion rate is mostly effected by System 1 processing.

 

Going back to media motives while spreading information, we can assume that the need to guarantee speedy and deep diffusion of information is forcing modern media to use System 1 distributed processing. While being successful in the acquisition of “likes” and “share”, this kind of information distribution is hampering collective wisdom first principle.

 

Speaking about the second principle: diversity and autonomy. Globalization is reducing diversity, while being all connected through social is also reducing the ability to process autonomous thinking. In one way or another, through social media sharing, we tend to aggregate out thinking to an already existing solution and the opinion of our social group is influencing our ability to remain different and unique in our decision making process. It is like we are loosing our individual autonomy in favor of an already formed social tribe opinion.

 

Last but not least, the aggregating function that enables the aggregation of distributed information processing into a collectively sound solution is also hindered by the inability of modern democratic politics to think with long term perspective and being instead completely focus on the “need to be reelected” short term perspective. This has always been the crucial distinction between statesmen and politicians. Statesmen should be able to think about future generations and rule accordingly, Politicians only think about next election and doing so miss completely out on the long term sustainability of their ruling.

 

Getting back to COVID19 issue, the collective delirium that is hitting the world population can be seen as a perfect case study for the ongoing failure of crowd wisdom principles in the social media information diffusion process.

 

Most of the distributed information has been tagged with System 1 effecting opinions, pushing on people most embedded biases of fear and uncertainty avoidance.

 

Decentralized information processing agents (world population) lost the ability to exercise differentiated and autonomous information processing, becoming a simple sound board amplifier of already packed diffused fear perceptions.

 

Governments have failed to provide any aggregating rational (System 2) function to the distributed fear processing, leveraging exponentially on the panic diffusion.

 

As always, when dealing with complex problems, we do not have a pocket solution to the ongoing phenomenon, but we can clearly state that the present information processing through social media is effective negatively the ability of the crowd to act smart. Social media communication is making the crowd dumber instead of smarter.