Like SortaGrey, I'm not a fan of carbon steerers (carbon blades are OK though). My reasoning is similar to his also. When considering the safety of anything I use an analysis that factors both the likelihood and consequences of failure. Where the consequences aren't likely to be serious, I'll accept a high rate of failure, however as the consequences become more severe I try to bring the likelihood to as close to zero as possible.

Given that a fork failure at the base of the steerer is probably the most catastrophic of all possible structure falures on a bike, I stay with the proven reliability of butted steel steerers for personal use.

That said, not all carbon fork failures are the same. Some manufacturers employ a safety net approach, usually using a few layers of fiberglass weave at the crown. This has enough flex to tolerate the bending that shatters carbon fiber and can hold the fork together a while. It's an old tried and proven approach going back to the use of wooden dowels in the bases of steel steerers.

I suspect that the biggest issue with carbon forks is that of sourcing and quality control. While I trust that the reputable firms have decent quality control standards, and choose and monitor their vendors carefully, I'm not sure this applies to all, and especially unknown suppliers. Then there's also the lack of long term history about product life under various conditions. If a given fork is OK for 50,000 miles of road use, how many of those can be on Pave?, How hard an impact from a deep pothole, unseen bump, or light impact can the fork tolerate?, if within the limit, how many severe jolts is too many?

When I ride, and descend at 40+ mph, I don't want to be thinking about my equipment. I ride for pleasure and trust in my equipment ranks much higher than any possible weight savings ever could.