I am sympathetic to the argument but disagree. I see it as a requirement that the code says must be met. I do not see it as the code saying, if it is in the pipe, or it may not be in the pipe, unless sized per table rating.
I see the strict reading saying the requirement is met by the EMT. The intent of complying, if the wire is intended as a required fault clearing path (and sold to the customer as a product that meets code) then the wire must meet the table rating. If the installation results in an additional unintended parallel path, like running an unrelated 12/2 MC cable parallel but happens to ground to the exact same two endpoints, I do not see anything that would prohibit or require removal.
There was an old code provision in the 1980's that prohibited "haphazard grounding" but I no longer see that. The closest to that would be 250.6, objectionable current on the grounding path.
In the late 80's I was tasked to install an "area protection monitor" that was a 20 amp circuit from each electrical closet (stacked six floors) that was monitored to see if the panel was live or lost power. I omitted the #12 green from the 3/4" EMT and the foreman threw a fit.
I cited the provision against haphazard grounding as the fault current needed to go to the source MDP over the panel feeders and not through the area protection monitor system which were in physically separate areas. The #12 copper could actually be a shorter run and lower resistance compared to the feeder copper EGC because the switchgear room was hundreds of feet away.
The foreman threw another fit when I told him this, that it was omitted intentionally.
More recently I cited omission of the GEC per 250.30 A(3) and A(4) plus 250.6 where the unlicensed people were claiming the 3/0 in the feeder was the GEC, when in fact it did not terminate on the grounding electrode or a busbar with a jumper to the grounding electrode. Feeder was for a 225 kVA delta primary in the PDU in the data center. In fact, the terminating point to satisfy (A)4 was less than 30 ft away already in the raised floor but the management company insisted on sticking to the lie. One end did not terminate on the grounding electrode or its group busbar, the other end did not terminate "at the same point as the system bonding jumper" but somehow the unlicensed people claimed it was the grounding electrode conductor (in the same pipe as the feeder). Their attorney denies wrongdoing !?