There are two answers, similar to what I said before. One is that we have knowledge from natural systems that CO2 is stored underground for time scales of, say, 100 million years, and there are many such deposits. Both that and the finding of oil and gas indicates that there are caprocks—formations that prevent the upward movement of CO2—that are widely found around sedimentary basins around the world, and they're capable of being secure enough for 100-million-year time scales.
That doesn't tell you anything about the engineered risks from the wells. The reason I told you about natural gas storage is that it bears on the engineered risk.
There's one more really important thing to say, which is that if you put CO2 in a deep saline formation underground, the CO2 dissolves in that “pore water”. At that point, instead of trying to work its way up, it's actually negatively buoyant, trying to go down. At that point, most people would say the risk is effectively zero. So after you put CO2 underground it actually gets safer with time, because the original pressure you put it under gradually dissipates and the CO2 gradually dissolves in the water, which means it's trying to go down and not up.
An important thing to say about this technology, unlike some of the other underground storage technologies, is that in general nature works for you, so that as time goes on beyond the initial period, it gets safer, not less safe. That's not to say that if we do this all over the world we're not going to have accidents; we will.