No, you use a sausage!

The trouble is you then have to replace bits and reset it, but then the only way you can test to see if it is now safe, is to trigger it again (with another sausage).
Safety systems are typically constructed of multiple elements, at the most simple an input ie system/instrument that detects the unsafe condition, a control system which determines if the unsafe condition passes a certain set of condition, and an output which is triggered to take some executive action. Each of these element will have a PFD (probability of failure on demand) and combined they will result in the overall PFD of the safety system.
When design a safety system you will have a target PFD, for example you want the system to target 1 or fewer injurie in 100yrs of using the saw. Some analysis would provide you with the initiating frequency, ie saw use data indicates 1 incident per year of saw operation. So your safety system needs to function correctly 99/100 times or a target PFD of 1/100. The PFD of each element can be managed separately and tested for independently.
On SawStop there is a continual current running through the system which is used as the input, if this current isn't present ie the cartridge is not installed or the input fails either the saw won't run or the stop system operates with a false positive. As a result an unrevealed failure of the input is not possible and the input does not require periodic testing.
The control system can be configured to self test on every start-up. Inputting a false input signal and measuring for the correct output signal. If this test does not pass then the saw doesn't start. This continual testing would again result in a very low PFD. Additionally this control system could be part of the cartridge so that it is replaced on each system firing.
The output element of the system (the explosive charge that detonates on demand) obviously cannot be tested fully. However the system could again send a much reduced current through the firing pin (or similar) to test there is a continuous path and that the firing circuit is functional.
The only thing you are then left with is the potential failure for the explosives to ignite, if the saw had gotten totally soaked/immersed this may happen but the same failure cause would render all the other control systems and the saw as non functional.
You cannot demonstrate any system or thing is 100% safe, and nothing is. However, engineering safety design and failure is a well understood discipline that designs and manages risks many orders of magnitude greater than loosing a finger. I'm quite comfortable that these principals will have designed a system that has a PFD sufficiently low that any individual user need not fret about the system not working when needed.
Sorry a bit of an essay but may interest a few folks on some of the rudimental elements of safety system design.
Fitz.