>People are really bad at risk
I don't think this is the right way to characterize the issue.
There are a huge number of unlikely disasters, and we round the small probabilities to zero.
This should not be called an error or a fallacy unless you can point to an alternative way of evaluating things which would be a practical improvement.
The idea that people should obviously take small risks seriously reminds me of Pascal's wager.
What's "obvious" at first is not so obvious once you consider the unlimited number of possibilities.
If you do assign a non-zero probability to any of them, then the total will be way too much, and you can't use it for making coherent decisions.