It really is a failure of regulation that these devices are still advertised and sold when they simply do not work.
For non diabetics whose BG stays within a tight range, of course a random value from within that range won't be very far off the true one, but it appears there's no actual correlation at all (which makes sense, because the claimed technology doesn't exist yet and nobody has ever demonstrated such a device working in real tests).
For anyone with more variable BG for whom such a device would be useful if it worked, paying attention to these faked values could be seriously harmful. For instance, someone with undiagnosed T2 could delay diagnosis and treatment because of the false reassurance from the watch that their sugars were normal.
For someone on insulin it would be disastrous.