Why are we so bad at assessing risk?
In a world filled with risks, how can we develop a better understanding of how we assess those risks?
Risk is all around us in the modern world, whether it comes from military threats, pandemics, artificial intelligence, climate change or economic instability.
Our responses to risk, however, often appear inconsistent. Some risks prompt immediate mobilisation, while others accumulate quietly for decades.
Human psychology evolved in environments where threats – predators, scarcity, conflict – were visible and immediate. Our brains are tuned to respond to sudden danger, but we struggle more with abstract, slow-moving or probabilistic threats.
Political systems amplify this bias, as democracies respond to electoral cycles, the media amplifies novelty and markets react to short-term signals. The result is an uneven risk landscape.
Visible, dramatic threats receive attention, even if they’re statistically rare. Meanwhile, slow, structural risks – environmental degradation, demographic shifts, infrastructure decay – struggle to maintain our focus.
Assessing longer-term risk
One obvious illustration is climate risk. Global insured losses from natural catastrophes have exceeded $100bn annually in seven of the past 10 years, according to the reinsurance group Swiss Re. However, political urgency around climate policy still tends to spike during extreme events and fade once the headlines move on.
Behavioural psychologists Daniel Kahneman and Amos Tversky’s work on prospect theory demonstrates that humans do not assess risk purely rationally: we prioritise immediate, vivid threats over abstract ones.
The public understanding of risk is important and statistician David Spiegelhalter, in The Art of Uncertainty,argues that clearer communication about probability, magnitude and timescale can significantly help with this. Unfortunately, institutional incentives often reward either alarm or oversimplification, because calibrated uncertainty is harder to communicate than certainty or fear.
Insurance markets provide a partial counterweight. When premiums rise, it reflects accumulated data rather than headline cycles – but even here, political pressure can distort pricing.
Risk isn’t only about probability: it’s about tolerance. What level of uncertainty is acceptable? And who bears the cost if preparation fails?
In recent decades, the drive for efficiency has outweighed the need for redundancy, so supply chains were optimised and buffers were reduced. This lowered costs in stable periods, but has increased our vulnerability to shocks.
Economist Nassim Nicholas Taleb has argued that systems optimised for efficiency become fragile. Just-in-time supply chains perform brilliantly in calm conditions, but fail abruptly when disruption arrives. The Covid pandemic exposed this tension, as have extreme weather events and geopolitical disruptions.
Human perception of risk
But risk perception is not only shaped by systems: it’s also shaped by biology.
Neuroscience shows that the prefrontal cortex – the region responsible for impulse control and long-term planning – continues developing into our mid-twenties. The brain’s reward systems mature earlier, however, a sequencing that helps explain why younger people are statistically more prone to risk-taking behaviour.
But developmental timing is only part of the story.
Addiction research has increasingly shown that reward-seeking behaviour is not simply about poor choices. It is deeply tied to the brain’s regulation of stress and emotional pain.
Gabor Maté, whose work explores the relationship between trauma, stress and addiction, argues that addictive behaviours are less about substances themselves than about attempts to soothe distress. In The Myth of Normal, he describes addiction as an attempt to regulate dysregulated nervous systems – a coping mechanism rather than a moral failure.
Public health data reflects the scale of the issue. The NHS estimates that hundreds of thousands of people in England experience problem gambling, while millions more are considered at risk of harm from alcohol, drugs or behavioural addictions.
Modern neuroscience increasingly supports parts of that picture. Studies of dopamine signalling show that intermittent reward – unpredictable reinforcement – powerfully activates the brain’s motivational circuits. This is the mechanism exploited by gambling machines, social media feeds and certain digital products.
The scale of that interaction is visible in the gambling industry itself. In the UK, gambling generates around £15bn in gross gambling yield each year, according to the Gambling Commission, with online platforms designed to encourage rapid and repeated engagement.
When rewards arrive unpredictably, the brain becomes more attentive, not less, and the design of these systems interacts directly with human neurobiology.
Trauma and risk tolerance
There is also a growing body of research in epigenetics suggesting that early-life stress can shape how individuals respond to risk later in life. Research by scientists such as Michael Meaney and Rachel Yehuda has shown that traumatic environments can alter gene expression associated with stress responses. These changes do not modify DNA itself, but they influence how genes are activated.
Risk tolerance, impulse control and stress reactivity are therefore not evenly distributed across populations: they’re shaped by developmental context.
This complicates the idea that individuals simply make rational calculations about risk.
Sam Conniff, in Fluke, explains how humans interpret uncertainty. He says that we tend to attribute outcomes to skill or intent rather than randomness and prefer stories of control to explanations involving probability.
That bias affects financial markets, entrepreneurship, politics – and everyday decision-making.
The challenge for us is not eliminating risk – which is impossible in complex societies, anyway – but recognising where our perception diverges from our exposure. Overestimating certain risks can lead to disproportionate spending: underestimating others can produce cascading consequences.
A calmer approach to risk doesn’t require alarmism. It merely requires calibration. Which risks are rising quietly? Which are politically convenient to emphasise? Which are economically uncomfortable to confront?
As the pace of technological and environmental change accelerates, our ability to understand and price risk accurately becomes a core competence. We need to learn that when our ancient instincts meet modern complexity, the mismatch explains much of our confusion.