I think you're misusing "terminal values" here. Risk aversion is a pretty stable feature of people's revealed preferences, so I can sort of see why you're putting it into that category; but I'd characterize it more as a heuristic people use in the absence of good intuitions about long-term behavior than as a terminal value in its own right. If you offered people the chance to become less risk-averse in a context where that demonstrably leads to improved outcomes relative to other preferences, I'd bet they'd take it, and be reflectively consistent in doing so. People try to do that in weak ways all the time, in fact; it's a fixture of self-help literature.
Well the reason I'd call it a terminal value is that if you asked people whether they would save 50 lives with 100% probability or 100 lives with 50% probability, people would tend to pick the former. When pressed why, they wouldn't really have an explanation, other than that they value not taking risks.
When I was a graduate student at the University of Notre Dame, I received a monthly living stipend of roughly $1,600. I decided to commit to giving ~10% of it to charity, and I had read in Peter Singer's book The Life You Can Save that one of the most efficient charities out there was Population Services International (PSI). Singer reported that GiveWell, a leading charity rating organization, had estimated that PSI's efforts saved lives at a cost of $650-$1000 each (pp. 88-89). So, I set up a recurring monthly donation of $160 to PSI, and kept it up for 15 months, for a total donation of $2,400.
I've been meaning to post the above information publicly for awhile, but was pushed over the edge by reading one of Eliezer's posts from a couple years back, Why Our Kind Can't Cooperate:
Since Eliezer's post is about rationalists, he stresses the issue of what arguments people voice. However, we know that just telling other people that you've given to charity makes them more likely to give. This is a point that Singer himself has emphasized.
I propose a thread for people to publicize their charitable donations. In light of the above, I'm especially interested to hear from people who've donated to the Singularity Institute for Artificial Intelligence. Once I acquire a regular source of income again in March, I intend to continue to primarily direct my charitable giving towards PSI, but maybe someone in this thread will persuade me to start giving to the Singularity Institute.