Their study, Democracy in the Age of Algorithms, says a significant reason is the way the human brain processes information: when new facts conflict with existing beliefs, people may reinterpret or reject them rather than change their views.

“For decades, the prevailing belief was that the root cause of mistaken views was primarily a lack of knowledge”, Anna Knorr, PhD, and Professor Małgorzata Kossowska from Jagiellonian University's Behaviour in Crisis Lab write in their paper. Psychologists call this the information deficit model.

Modern democracy is based on the premise that when citizens have access to reliable information, they can make better decisions – for example, by better evaluating the actions of those in power, understanding the effects of proposed changes and making informed decisions about their representatives. Access to information should also allow people to reconsider their beliefs and change their minds.

However, the researchers say reality is more complex. Numerous studies show that although disinformation can deepen social divisions, simply correcting false information rarely leads to a lasting change in beliefs. Exposure to well-documented facts can sometimes even reinforce existing opinions.

Two people can read the same article, watch the same material or consult the same scientific report and still reach different conclusions.

The digital revolution has made access to information easier, faster and more widespread than ever before. But more information does not necessarily translate into a better understanding of reality, the researchers say.

“Although we have access to vast amounts of knowledge every day, reaching shared conclusions has become increasingly difficult. Information increasingly fails to fulfil its basic function - instead of helping us better understand reality, it merely confirms what we already believe”, they write.

In other words, confronting people with facts through fact-checking rarely leads to lasting changes in their views. When facts contradict existing beliefs, they can sometimes be interpreted in ways that reinforce those beliefs.

According to the researchers, the explanation lies not only in the way digital platforms and social media operate, but primarily in the way the human brain works.

A key concept for understanding this process is predictive coding. According to Knorr and Kossowska, the brain does not passively perceive the world. Instead, it constantly generates predictions based on prior knowledge and compares them with incoming information.

When new information aligns with existing expectations, beliefs become stronger. When it contradicts them, the brain must choose between changing its model of the world and reinterpreting or rejecting the inconvenient fact.

The stronger a prior belief, the more difficult it can be to change. The strength of such beliefs is also largely shaped by the information people have encountered previously, the researchers say.

The mechanisms governing cognition can be further strengthened by the modern digital environment.

Recommendation systems on digital platforms are designed to keep users' attention focused on the screen for as long as possible, which translates into advertising revenue.

Algorithms learn users' weaknesses. Content that generates strong emotions such as outrage, fear or anger can spread faster than calm, balanced explanations. As a result, users are increasingly exposed to messages that confirm their existing views and less often to information that could challenge them.

At the same time, the researchers say, people are not merely passive victims of this process. Under the concept of active inference, the human mind actively seeks environments in which its predictions are confirmed, while algorithms can learn to exploit this tendency.

The researchers say this is not a marginal issue. Three-quarters of Europeans, particularly young people, now primarily obtain information about the world from the internet, where digital platforms compete for their attention.

Instead of fostering a shared understanding of reality, the architecture of these platforms can reinforce existing beliefs, deepen social divisions and influence what people consider to be true, according to the researchers.

The authors say democracy does not need a society that agrees on everything. It needs a society capable of disagreeing on the basis of shared, commonly accepted facts.

When citizens operate in different information environments, informed debate and decisions on common issues become more difficult, they say.

According to the researchers, the challenge is therefore not only to combat disinformation but also to build an information environment that encourages people to encounter different perspectives and critically evaluate information.

They say today's information landscape is not the inevitable result of technological development, but the result of specific design, business and regulatory decisions. This means it can be designed differently.

The researchers say European regulations, including the Digital Services Act and the AI Act, show that the operation of recommendation systems has also become a subject of political and ethical debate.

“At the same time, regulation alone will not be enough. A better understanding of the mechanisms that govern human cognition is also required. Knowledge from cognitive psychology and neuroscience should become one of the foundations for the design of future digital environments”, Kossowska and Knorr say. (PAP)

PAP - Science in Poland

jjj/ zan/

tr. RL