Risk compensation is irony encapsulated; the more measures that are taken to ensure people's safety, the more dangerously people will act to make up for it. Perhaps people are just designed to seek a certain level of risk.
And yet, safety measures are good things. Sure, they can be overdone, but everyone likes their staircases to have railings, and everyone likes their cars to have good brakes, and everyone prefers going home at night to break open a bag of potato chips to spending time in the hospital. People make mistakes as they go about their daily lives, so minor precautions to make sure that those mistakes don't have disastrous results is basic kindness and common sense. The problem is, those precautions that are taken with the idea that the mistakes they compensate for won't get more frequent or more egregious. Sadly, this is not the case.
It seems that people engage in risk compensation; becoming more careless as the consequences for that carelessness become less. This kjnd of compensation is often seen in ordinary life. No one carries a half-full cup of lukewarm water the same way they'd carry a cup of scalding coffee filled to the brim. The problem is, risk compensation is also used in extraordinary and dangerous situations. For example, the rate of skydiving fatalities stayed the same for decades after the eighties despite the increasing number of failsafes and technological improvements to the chutes. Once just jumping out of a plane wasn't enough for people, they began showing off, doing maneuvers that killed them despite their functioning parachutes.
Even people who don't throw themselves out of planes regularly take their life in their hands. Driving is a subject of much study in the area of risk compensation. A huge number of people drive, or are driven, and although it can result in devastating consequences, it's done often enough that it feels routine, safe, and boring. Seatbelts are a good idea for drivers, but can be a bad idea for pedestrians and bikers. An experiment in which people drove go-karts showed that those who switched from an unencumbered ride to being strapped in safely responded by increasing their speed. Conversely, when Sweden switched the side of the road that people drove on - something that might lead to more accidents by confused or distracted motorists - the accident rate dropped. It only crept up again when drivers started to become accustomed to the change.
One of the main pieces of ammunition for proponents of risk compensation is a survey of British drivers. Truck and cab drivers readily admitted that if their vehicles were loaded with explosives they would drive more slowly. Gordon Tullock, an economist, once joked that if the government wanted people to drive safely, they'd mandate a spike in the middle of each steering wheel. Of course such a thing would never happen — nor should it, morally — but ensuring the person most in control of a situation will be damaged by their mistakes can lead to much safer behavior than ensuring that they'll be protected, even if others won't be. Everyone's careful not to incite a revolution if they know they'll be the first up against the wall.