One of the risks is the fuel getting jammed. Since they're spheres, they should have very low friction. But they already saw that defects in the coating can raise that risk. They would need very strict QC on manufacturing the pellets, and the entire system must be designed to mitigate the chance of wear causing damage. There would naturally be a buildup of debris over time, but fine carbon dust usually serves as a lubricant anyway. They would need to prevent contaminants entering the core.

Even if there was a jam, is there a foolproof way to stop the input, even during a power failure? Can the pellets sit in the reactor forever without getting too hot when the cooling is down?

Is any of this human controlled? Part of Chernobyl was someone ignoring a failure and choosing not to shut down until it was too late - is that a possibility here?

So yeah, saying failure is impossible is literally what they said with the Chernobyl-style reactors when they were new. They did safety tests on those to see what would happen if the power failed, which was itself the catalyst for the failure. Just say that you have a new, extremely safe design, be open about how it works, and don't tempt fate?