Abstract
Collective stupidity is fairly common in organisations. Firms full of very smart individuals who act in reckless ways that destroy them. Governments and nations that engage in acts of self-harm, descending into economic decline or civil war. Armies that repeat failed tactics.
It’s easy to see collective stupidity as something others slip into because of their own idiocy or moral failings. But this perspective misses much. We are all party to such follies.
Here I define collective stupidity as the mirror of collective intelligence. Collective intelligence emerges when the combination of individual intelligences produces thought and actions that are more intelligent than individuals alone. Collective stupidity is the opposite, emerging when thoughts and actions are less intelligent than those of individuals alone (and are seen as such in retrospect).
I draw on my experience working in many governments, civil society and now in academia, to argue that often collective stupidity in organisations is in part an emergent property of social dynamics. Adam Smith’s theory of the invisible hand showed how self-interested actions aggregated into an effect that benefitted all. The pursuit of individual interest contributed to the common interest. The patterns I describe are opposite in nature: they aggregate rational, pro-social individual actions into results that work against the common interest. The reasons why we are so often part of collective stupidities is not that we are stupid, but rather that we are well intentioned.
The two sources of collective stupidity
There are many kinds of collective stupidity – here I focus on cognitive failures that in retrospect are considered by the organisation itself, or its wider community, as having led to unnecessary errors.
Much of the recent theory of intelligence focuses on error – and the ability to update in response to a gap between model predictions and input data. The ability to do this well is one definition of intelligence in an organisation (and updating can be within a model, involve new categories or even new cognitive systems). But there are powerful forces and tendencies which get in the way of correcting errors and these offer clues to the persistence of collective stupidity.
The first is the tendency to conserve cognitive energy. Intelligence of any kind is costly. It demands energy and attention. This explains why we use everyday heuristics or rules of thumb that generally work well and don’t require too much thought.
The second relevant pattern is the tendency to maximise positive social feedback and belonging. We all want to be liked and accepted. We are social animals and depend greatly on feedback from others that tells us we belong. Like cognitive energy conservation, maximising positive social feedback and belonging is rational, as it helps us survive and thrive.
As a result, we try to align how we think with the thought of others around us, to see the world in similar ways and to act mimetically, copying others. Rituals, signs, brands, symbols and jargons all contribute to this.
These tendencies – the combination of cognitive energy conservation and social feedback maximisation at an individual level – aggregate to help align social cognition, which is often a strength of human groups. But they also feed collective stupidity when not mitigated.
Some drivers of socially amplified stupidity
Collective stupidity can be amplified by our pro-social tendencies: our deeply ingrained desire to fit in and be accepted. These are some of the main tendencies which fuel stupidity within organisations.
Group commitment
Evolution has shaped us to become part of groups and much social psychology documents how this happens. For example, we recall information in different ways if it comes from within groups we are part of rather than from outside. We tend to be much more upset by in-group argument than arguments involving out-groups. And our bodies as well as our minds notice and struggle with group differences, even when these are arbitrary and trivial. This ready experience of being in groups can then be extended to other groups and identities.
Loyalty to the group and peers
All groups depend on loyalty – loyalty to each other, and loyalty to the group to help it through inevitable setbacks and defeats. Loyalty matters in every relationship that has any meaning in our lives. We expect shared experiences, love and friendship to count for something, particularly when times are hard, and choices are not easy. Doing things together creates a credit that we expect to be able to draw down, and in this we are helped because familiarity makes us partial to our own family, friends or nation. But loyalty again often overshoots, preventing people from sharing vital information or voicing concerns.
Shared identity
The related factor is that groups build up an identity – a sense of shared self, that comes with stories, symbols, histories, dress codes and more. These are functional – they make it easier to cooperate and collaborate. They encourage mutual help and sacrifice. But they also often overshoot when others are seen as subhuman, alien or unworthy of help.
Suggestibility and conformity
Humans are very suggestible – it is one of our great powers and enables us to learn quickly and to cooperate. But this suggestibility in groups also amplifies error as well as accurate understanding: panics, bubbles, bandwagons and collective hysteria feed off suggestibility. We are often willing to suspend our own judgement and reason if we believe others see things differently (Asch, 1987).
Harmony and homeostasis
Many things seek equilibrium and that includes human groups. Just as our bodies respond to divergences from the norm by trying to return us to a steady temperature, or blood sugar level, so do groups tend towards equilibrium and harmony. Conflict and divergence are costly and uncomfortable. They take time and effort. And so most of the time we suppress thoughts that might lead to conflict. Often that is for the best. If every thought was spoken, everyday social interaction would become unbearable. But for groups to think well they often do need diverse and conflicting views: alternative perspectives on what the facts are, how they should be interpreted or what options might follow.
Excessive respect for overconfidence
The world is vastly more complex than our brains can grasp. But many have too much confidence in their own abilities, and this often gives them status with others who are even more uncertain than they are. Overconfidence is rewarded socially, and in most groups people defer excessively to the overconfident and the charismatic. This can make the group or organisation act in foolish ways (some of the current leadership in the USA arguably exemplify this point). Organisations often reward and promote the overconfident who reinforce the legitimacy of the organisation – but precisely for that reason they are more prone to error unless mitigating processes are put in place to test, challenge and interrogate the overconfident.
Socially reinforced errors
These various social tendencies, which can be found in almost any organisation, help to fuel some of the common patterns of error and collective stupidity, drawing on the interaction of cognitive energy conservation and social feedback maximisation – where social dynamics make it harder to observe, interpret and think. Here I sketch a few of the main patterns.
Conformism and groupthink
Groups tend to groupthink and loyalty and identity encourage conformity. There is strong evidence that in meetings people often edit or suppress the most relevant information they could be offering if they think it will in any way undermine the coherence of the group, or their own social standing.
Oblivious consensus
If we spend time with people like us, as most people do, we tend to believe that their beliefs are much more widespread than they really are. This pattern is greatly amplified by social media, but it’s been common throughout human history. As a result, we misinterpret local consensus as general consensus, and that local consensus may often include stark failures to see the big picture.
Collective forgetting and editing out
Commitment to the group, loyalty and identity help people ignore potentially relevant information, and these myopias are reinforced socially: if we see others ignoring them we are much more likely to do so ourselves. Groups are held together both by what they remember and by what they choose to forget. Remembering too much about your errors, moral failings and wrongdoings is not good for group cohesion. Some selective amnesia is healthy. But too much is unhealthy and distorted memories then damage the ability of the group to thrive; for example, remembering a defeat as the result of betrayal rather than errors on its own part.
Collective distractions
All individuals are prone to being distracted, but social interaction amplifies these distractions greatly, again crowding out reflection. Karl Marx called religion the opium of the people, in part because its rituals were distractions. Today, social media have greatly strengthened the social feedback around distractions – so that the individual tendency to cognitive distraction is amplified by seeing others distracted in the same way. Bubbles and echo chambers encourage distraction as well as many of the other pathologies described here.
Collective mental illness
Another aspect of these problems is collective mental ill-health (Mulgan, 2021), where there are some parallels between widely shared disorders and the kind of individual psychological disorders documented in the psychology manual, Diagnostic and Statistical Manual of Mental Disorders (DSM-5). These examples of collective mental ill-health could include delusions, depressions, obsessive compulsive behaviour and post-traumatic stress disorder (PTSD). These appear to be partly contagious within groups and they can certainly lead to maladaptive behaviour. This certainly warrants more research, especially as organisations are collecting far more data on the mental well-being of their employees.
Countering collective stupidity through better collective metacognition
Most of the types of collective stupidity I’ve described above result from the interaction of energy conservation and social feedback maximisation. An interesting feature is that they often feel good at the time – they are experienced as pleasurable and life enhancing even if they are sowing the seeds of disaster, though other contributors like tiredness and stress do not. (Heffernan, 2012).
But many institutions and groups have learned how to counter these trends. They partly do so through education, training individuals to spot false reasoning and common patterns of error. But they also do so through metacognition, spending resources on methods of thought that protect from the traps of collective stupidity, which some researchers have described as ‘functional stupidity’ (Alvesson & Spicer, 2012).
First, they disaggregate cognitive processes – dividing roles up between different people or groups, tasks and times: chunking it up. All organisations manage cognitive divisions of labour; those that involve everyone in every decision rarely thrive.
Second, they orchestrate triangulation and doubt, to provide alternative perspectives. They force attention to meta questions and metacognition; how would I know if this wasn’t true? What would someone else think? What might I think in 20 years’ time looking back? Red teams are a good example of this.
Third, they organise loops – looping back repeatedly to question, think, debate, challenge: this is very common in the best teams in sports, groups or orchestras. In a team, a group or a community argument can be painful but it allows the group to reach better solutions through exploration.
Fourth, they sometimes deliberately slow down processes to improve cognition, allowing more time for reflection, detachment and distance. Crowds think fast, often with disastrous effects – rushing to judgement on guilt, or what actions to take. Other processes are deliberately slower, including elections but also the workings of courts, boards and other institutions. Indeed, these often deliberately slow down processes of thought the better to ensure wise judgements – separating diagnosis from prescription, or judgements of guilt from punishments.
Fifth, they help people understand what others really think, countering the tendencies to assume that others think just like us, or conversely the assumption that strangers have much darker intentions than they really do. Understanding where others are coming from, their feelings, reasoning, hopes and fears can be crucial to avoiding foolish actions.
Sixth, they take care to achieve a healthy balance of diversity and integration. Diversity of all kinds can improve performance, as documented by Scott Page and others. But some kinds of diversity do the opposite, feeding conflict, unwillingness to share and mutual incomprehension, with people talking past each other. Often, decision-makers lack the intellectual tools to synthesise diverse perspectives. Healthy diversity requires both variety of views, backgrounds and cognitive styles, and a capacity to integrate that is helped by shared languages, frames and ways of thinking.
Finally, successful groups listen to unwelcome information, and unwelcome ideas, rewarding dissent and whistleblowing. Sometimes that involves institutionalising honest reporting (as airlines do for problems and near-misses). Sometimes it involves rewarding mavericks: the scientist heretic or lone voice who may be wrong most of the time but may be right occasionally. The collective interest depends on giving them some space to think and argue.
This essay has suggested a general theory of collective stupidity, arguing that it is fuelled by two rational dynamics – the minimisation of cognitive energy and the maximisation of positive social feedback. These explain many of the common phenomena of conformism and error. But I’ve also shown that some organisations systematically counter these dynamics through mobilising metacognition.
This could be a useful avenue for both research and action, building on recent work on collective intelligence and intelligence at the level of whole systems (Friston et al., 2024). The paradox of the modern world is that we have far more educated populations and far more access to smart technologies – and yet so often are collectively foolish. This approach may offer some clues.