The society of widespread lies

On the same topic:

Do we live in a society of lies and deception? And if so, with what consequences? Some contribution to the debate raised by such questions can come from comparative ethology and evolutionary biology. Deception, although often interpreted as pure selfish cunning, actually shows an intricate web of effects that go beyond the single individual, touching on group dynamics, social regulations and even cultural transmission.

Lying, from an evolutionary perspective, takes shape in a system of costs and benefits that has shaped communication strategies since the dawn of our common ancestors with other primates. In an environment where resources – food, reproductive partners, social allies – are scarce and unevenly distributed, the ability to manipulate the beliefs of others can offer an immediate selective advantage: an individual who manages to conceal a resource or present false information increases its reproductive fitness if it is able to transmit its genes more effectively than its rivals . However, for this strategy to stabilize in the population, the individual gain must exceed not only the energetic costs of deception (for example, the cognitive expenditure in planning the deception), but also the social costs of discovery, such as loss of status or exclusion from the group.

According to game theory models, two individuals who lie and then discover each other are both at a disadvantage compared to a scenario of trust-based cooperation. This results in a form of evolutionary equilibrium in which occasional tactical lies reign, limited by mechanisms of punishment and reputation. If the detection rate of lies is sufficiently high and the punishments severe enough (e.g., isolation, reduction of social support, loss of reproductive opportunities), deception is no longer profitable in most interactions, and only under particular conditions – such as high-grade asymmetric competition – does it emerge stably .

This dynamic equilibrium quickly makes fraud costly even when it is not discovered, because to be credible it must use significant cognitive and physical resources, which leads to a decrease in liars – soon they are only those who have sufficient resources to invest and a high return on investment. This is the concept of “handicap signaling” introduced by Zahavi and formalized by Grafen, which clarifies how a deception can be evolutionarily advantageous only if it entails a cost for the liar that is higher than what the less gifted are able to bear. In animal communication systems, in fact, many signals are designed to be so expensive or easily disprovable as to make any attempt at bluff futile: in fact, a conventional truth of communication is established that protects the reliability of basic interactions. However, when brain volume and cognitive abilities reach high levels, as occurs in chimpanzees and even more so in humans, the price to pay for deception increases in proportion to the complexity of the deceptive message, which must pass the scrutiny of superior mental capacities. Thus, starting from simple tactical bluffs, the way has been opened to refined lies, capable of taking on elaborate forms in political speeches or marketing strategies, precisely because their cognitive and social cost has become sufficiently high to guarantee, at least in theory, a medium-high credibility constraint. Liars have become rarer, but lies much more effective.

Throughout the human lineage, then, selection has favored not only the ability to deceive, but also the ability to detect and punish lies: a cognitive arms race has pushed us to develop theory of mind and metacognition, tools with which to assess the honesty of others. Individuals who were better at detecting lies were more reliable as cooperative partners, and thus enjoyed advantages in mutual support and parental care. Likewise, those who could deceive well, without being detected, obtained extra resources; hence the coevolution of lying and detection, driven by a dynamic of bilateral selection.

A further level of complexity emerges if we consider lying from a group selection perspective. Groups in which lies were kept within reasonable limits – thanks to social norms and sanctions – tended to cooperate better, to consolidate and to compete successfully against groups in which deception was endemic and disintegrating. This mechanism explains why, despite the individual gain offered by lies, their excessive diffusion was contained by cultural selection: communities with an intense “reputational cage” prospered, while those in which deception was uncontrolled disintegrated . Note well, however, that this mechanism keeps under control the lies that we could define as selfish: lies that, instead, function as a glue and increase cooperation – identity lies – are favored and impervious to the control mechanisms stated.

And so we come to the crucial point: if a lie favors the cooperation of a large group, it can guarantee both the reproductive success of the individual and that of the entire group. Identity lies of this type are conspiracy lies, and on a larger scale marketing lies that leverage the identification of consumers with very specific groups; all the way up to political lies, which work and are much more successful than the story of the facts they obscure.

Since in our species lying is eminently mediated by language, it is clear that the role of large-scale linguistic models (Llm) has now been added to the dynamics described . Llms have dramatically shifted the balance described, generating deceptive texts of superior quality, structured with coherence and persuasive style, and at an extreme speed and scale; in this way, the cost for the lie we were talking about earlier becomes very low, and the difference between those excellent liars able to pay it to gain an advantage and the average individual is eliminated – except perhaps for the cost associated with knowing how to use the tool well. Recent studies show that Llms are more convincing than human communicators in debates on sensitive topics, modulating arguments based on minimal demographic data to maximize the persuasive impact . Furthermore, “grooming” campaigns by hostile actors – which sow networks of false content intended to feed AI algorithms – have already demonstrated how the responses of conversational agents can be indirectly oriented, amplifying the spread of distorted narratives . On the one hand, these technologies transform deception from an occasional tactic to a weapon of mass influence, capable of eroding collective discernment without users realizing it; on the other, they make the cognitive cost of deception so low that it can multiply liars indefinitely. Under these conditions, it is easy to predict how the ethological and social balance that guarantees the cohesion of groups and even large human and animal populations, based on the cost of generating credible lies and the risk associated with the loss of credibility in the event of discovery, could collapse.

The drastic reduction in the “cost” of lying induced by large-scale linguistic models and digital platforms has transformed what was once a rarely used tactical signal into a widespread practice, capable of triggering waves of mass violence when shaped around collective identities . If until yesterday weaving a complex deception required time, cognitive effort and the risk of social sanctions, today a well-crafted prompt is enough to generate extremist speeches, ethnic or religious narratives using algorithms capable of refining tone and style based on the prejudices of a given group. This ease erodes the original “handicap signaling”: the cost associated with generating a credible lie, the cost that simultaneously makes it credible and rare, becomes zero.

In addition, the speed with which false and identity-based content replicates and adapts to local contexts excludes any possibility of spontaneous control: deceptive information, now unrestrained, becomes a weapon that pushes latent conflicts to explode with unprecedented violence, structuring communities of enemies of great breadth and great determination . Today, we call this politics; and how much it produces the same disintegrating effects, violence and wars, is now reminded not only by the data we have in the case of non-human primates, but directly by the contemporaneity we are living in.

Everyone's effort, therefore, must be directed in a precise direction: not only in recognizing and preserving the discussion of the facts, that is, of the evidence we have available about one thesis or another, but above all in preserving a collective and shared narrative that uses the rule of proof as its foundation. The battle is for the sharing of a method, not for this or that more or less well-founded hypothesis. And it is a battle whose stake is not only our ability to adapt to the physical world, impossible without the objective analysis of the data available, but the very survival of a complex society beyond the level of perpetually warring tribes and oligarchs with unlimited power.

More on these topics: