C2x: what is the rationale for alignas to be a keyword rather than a macro?

In 1990 P.J. Plauger wrote (emphasis added):

Standard C offers you an additional level of security, however. It is a level offered in no other language standard that I know. It promises that if you avoid certain sets of names, you will experience no collisions. Thus, Standard C makes it that much easier to write highly portable applications.

In C11 keyword _Alignas (for example) was introduced with the accompanying macro alignas defined in <stdalign.h>. Here we see that the keyword is _Alignas, not alignas (since in pre-C11 alignas is not reserved). Hence, there is no collision with possible user-defined alignas.

However, in C2x the alignas is a keyword and <stdalign.h> provides no content (and C2x says nothing about __alignas_is_defined macro -- defect?). It means that in C2x any pre-C2x code containing user-defined alignas will cause semantics violation and, hence, breaks backward compatibility.

Questions:

1. Does it mean that since C2x the "you will experience no collisions" does not hold any longer?
2. What is the rationale for alignas (for example) to be a keyword rather than a macro?

The proposal for these change (available at https://open-std.org/JTC1/SC22/WG14/www/docs/n2934.pdf) argues that the naming strategy in regards to new keywords has already been inconsistent with previous standard versions:

some were integrated using non-reserved names (const, inline) others were integrated in an underscore-capitalized form. For some of them, the use of the lower-case form then is ensured via a set of library header files.

Further using the same keyword naming as C++ (for compatibility purposes) is also mentioned in the proposal, since some keywords originated in that language and were later added to C.