I've heard this concern raised before, and the real question is whether the thermal expansion of the fuel salt acts as a strong enough negative coefficient of reactivity.