Quite so. AsI recently wrote, what is needed in such situations is a risk-benefit assessment/judgement/analysis - the outcome of which may vary very widely ....
In the case of seat belts, the probability of them doing harm is so low, and the probability of them doing good so high, that it is a no-brainer to conclude that, 'on balance', they are worth using. Much the same is probably true of "... crossing a road today/driving a car somewhere/ having a vaccination/ visit a GP surgery ".
However, the situation with bonding is not quite so clear cut. However, in many cases of 'unnecessary' bonding (certainly with some of the 'sillier' cases), it would seem that the probability of it 'doing good' is very small but, from the point-of-view of a person in contact with a live conductor/part, increasing the amount of 'unnecessarily earthed' touchable metal around (an inevitable consequence of some unnecessary bonding) will appreciably increase risk - so, on balance, the conclusion probably should be in favour of not installing unnecessary bonding (of otherwise unearthed things).