Safetyism and alignment problems
September 22, 2014 — November 14, 2024
Every time I come back to Australia, I notice how damn safe everything is. Sometimes this is great — I love not dying while crossing the road or getting mugged. Sometimes this feels smothering, as when required to get insurance for banal things like a one-off event, or filling out risk assessment forms for a picnic.
Thus: a notebook to wonder how to align intuitionistic ethics, intuitionistic risk perception and getting things done, specifically, attaining better balance between safety and risk.
Safetyism, if I’ve understood it correctly, describes the situation in which most people would argue that too much effort is spent on safety, to the point that it impairs other goals, or even becomes self-defeating, i.e. becomes more unsafe. At the government level, this is often denigrated as nanny state. Observationally, many people believe they are in such a situation.
My own preferences for how safe things should be, and how regulated, are idiosyncratic. So are yours; part of the complication here is that we need to find good balances of safety for society as a whole.
1 The role of inaction
A common theme […] is a willingness to kill hundreds of thousands of people through inaction, before decisionmakers are willing to risk taking any unpopular action.
The Copenhagen Interpretation of Ethics points to one possible explanation, but it shouldn’t be forgotten that leaders are to a large extent giving the public what they want in all of this — it’s just that the public has pathologically low standards and a bizarre level of change aversion.
Specifically,
The Copenhagen Interpretation of Ethics says that when you observe or interact with a problem in any way, you can be blamed for it. At the very least, you are to blame for not doing more. Even if you don’t make the problem worse, even if you make it slightly better, the ethical burden of the problem falls on you as soon as you observe it. In particular, if you interact with a problem and benefit from it, you are a complete monster.
2 Safety as a public good
TBD. See also squeaky wheel equilibria.
3 Counterproductive safetyism
e.g. Avoiding peanuts TBD.
See also comfort traps.
4 Safe spaces
Some people are concerned also with psychological safety, which leads naturally to the question of psychological safetyism. What is a good social norm for delineating and enforcing safe spaces? Contrariwise, how do we negotiate whose safety and in which space?
5 Anti-safetyism
What about when we spend too little effort on safety? As someone who worries about catastrophic risks, I am clearly concerned that we underinvest in safety of some kinds.
TODO.