In software, one more-or-less generally-accepted result is that, when you fix a big, the probability of introducing new bug is somewhere between 20 and 50%. One would hope that it's much less for critical software, but it's not zero.

In the same way, what's the probability of introducing a new defect when replacing the wiring harnesses? It's not zero. It's never zero.

So the answer to "should we fix this issue?" is not "yes, of course". It's "will the fix make things better, or worse?" And of course, you don't know with certainty, but you make your best estimate of the probability of the new problem being worse than the old, and then you decide.

So even if "eliminating risks to the best of our ability" is in fact the right goal, that doesn't always equate to "eliminate this risk", even if we can do so.

> I don't like the idea of manufacturers policing themselves.

Here I agree with you completely. We've seen that movie; let's not watch it again.