I would think so. The problem isn't that the tag contains a banned word, the problem that the source is one that OSM doesn't have permission to use.
osm.wiki/Copyright#Proprietary_data

If the feature has always been listed as sourced form Google, it should be deleted rather than have its history obscured.

This seems to be an odd conversation on a changeset that didn't change the name of the street.
P.S. the dual carriageway part by the roundabout still has the old name.

CC BY isn't considered compatible with OSM without a waiver so using the linked photo as a source is problematic:
https://blog.openstreetmap.org/2017/03/17/use-of-cc-by-data/

There doesn't seem to be a CC BY Waiver for this source at osm.wiki/Contributors as for other CC BY sources. Do you have a link to one?

Is this the correct end of the cay? There seems to be a prominent shadow towards the western end.

Have you surveyed this?
Piers seem unlikely for boundaries between salt pans and does not appear to match the imagery. Man_made=dyke might be more appropriate.

These edits don't seem to be in Haiti.

Did you mean to place traffic signals off the coast of Africa?

Why have you deleted the Nassau node?

Even within your preferred database this Great Guana Cay exists as UFI: -1543360.
Please do not destroy other people's work without checking.

FYI those admin boundaries are a bit approximate.

The new Royal Stingray Lagoon geometry is confusing a few tools (contains the same segment twice). Did you intend to transform it from a horseshoe shaped lake to a lake with an island?

https://www.lexico.com/en/definition/bight

natural=bay (paragraph 3)

If you are interested in widely accepted practice I suggest you read about the principles of One feature, one OSM element [1], which is at the root of this discussion, and tagging for the renderer [2] which you implied you were doing in your first response.

I don't think "90+% of the community" would accept scrub plastered over existing beaches and properties as an improvement nor would they support a change from the primary way to map islands in the sea [3] to a multipolygon that excludes large chunks of that island.

[1]: osm.wiki/One_feature,_one_OSM_element
[2]: osm.wiki/Tagging_for_the_renderer
[3]: place=island#Islands_in_the_sea

Both OSM Inspector area detection, and the link you posted (Figure 18 1st image) say the multipolygon you have made is wrong. The satellite imagery also shows that you have incorrectly extended "scrub" tagging over areas that are construction, beach, cleared for residences and bare rock. And you have decisively excluded areas that are clearly part of the island from the tagged island in order to create the false assertion that the island is all scrub when it blatantly isn't.

It doesn't matter to me if the island related tags are on the coastline polygon (simpler, more traditional) or on a pointless single-child "multi"polygon (also an OSM Inspector error), but what is currently tagged is nonsense.

Norman's Cay is the name of the island, not the scrub. The scrub is a separate entity. As the coastline way is relatively short it is probably the best host for the island related tags.
OSM Inspector (and probably other software) is also alarmed by the segment shared between inner and outer ways.

Is it not a little unconventional to exclude the airport area from an island multipolygon?

Please do not use descriptions as names.

Yes, I've been meaning to go through these and delete most of the 'descriptions'.