All joking aside .... I think that the author really does advocate placing GFCI protection in the panel - which would mean the end og the GFCI receptacle. It's safe to assume that the device makers will strongly oppose this effort.
I'm not sure how the 'self test' feature (present in all GFCI's made since 2003) can test for actual movement of the solenoid inside - which is what corroded it the failed examples, to the point where it could not move.
It should also be noted that the 2003 revision of GFCI standards also introduced some additional corrosion protection. The advent of "WR" devices introduced even more corrosion resistance.
In other words, the 'problem' has already been addressed. Now, it's a bit late; it would have been nice had we been aware of the issue back when it was a concern! After all, "corrosion resistant" devices have ALWAYS been available, though expensive, and usually special-order. We didn't buy them, as we didn't see the need for them. A study, such as the one cited by this author, would have made all the difference 'back then.'
Which is why I believe that the main purpose of this article is to drum up support for placing GFCI protection in the panel.
The short biography of the author included with the article is devoid of any connections with any manufacturer. That may very well be the case - though past antics of manufacturers (Hydrolevel vs. ASME) do justify a certain amount of paranoia. As it stands, where you place the GFCI protection is strictly a design choice.
Yet, I've seen these issues fester, often for decades. It's not going away, and we can expect this proposal to appear at EVERY code cycle from here on. One recent example was the campaigning on internet forums that led to 'multi-wire' circuits being essentially banned from residential work. Even in that example, the proposals didn't directly attack mwbc's so much as place so many restrictions on them as to make them unattractive.
You've been warned.
