I like the idea of a time-to-risk ratio. Perhaps the longer you spend developing something, the less risk there is of a problem, or the more you spend on it would work too, either more funding or more time.
And I didn't mean to make it 30% chance. I was just throwing out numbers to get the system across. I think if you do it the quickest and cheapest you can a 15% chance and if you do the carefullest way possible a 5% chance? Also, you could get upgrades that reduce the damage level, such as agencies regulating safety procedures. These could be on a seperate, streamlined tree, that instead of having lots of different paths just has a strait path for each dangerous technology.
MOD: Also, you should have to determine how many turns you give your scientists to develope the technologies (two turns min, maybe four max?), and there should be a safety cost if you shorten the deadline. So if I said I was going to give my people 4 turns to develope nuclear power, it would reduce the chance of problems from 15 to 5. Then the next turn I decide that I NEED that power ASAP, I could reduce the time to 2 turns, but I would not get the safety bonus for turn 3 or 4 and I would have a penalty of 5 or so.
Finally, I think the longer you have a technology, the smaller the risk should be. Maybe reduce the risk by .5% every turn you have the technology without problems, and don't reduce the risk on turns you do.