I would take it a bit further, though, and propose a massive frequency swap.
In the NARBA change in the early 40's, only a bit over 700 stations were affected. And the moves were, at most, 40 kHZ and many were 10 to 20 kHz.
To do what you propose means moving stations widely over the dial. The cost in rebuilding transmitter sites, often with new directionals, nearly always with total replacement of the transmitter and antenna tuning components would be enormously expensive.
Add in stations that would have to increase tower height that are on land that will not support guys for a taller structure or where zoning prohibits such change, and some stations will not survive.
Stations that need new directional designs might need a year to three years to get zoning, and have to spend hundreds of thousands on new land. Many would go silent.
Listenership would probably improve once the stations were stronger, and less prone to interference. Although the frequency swap nationwide would be massive, individual listeners would probably only have to remember a handful of new frequencies at most, which could be heavily promoted before the swap.
This does not change 30 years of cheap AM radios, the fact that AM at its best can't sound as good to listeners who did not grow up on the band and the still very limited coverage of most stations, and it won't work.
75% of Americans never use AM. Those that do are predominantly over 55. Nobody wants to invest in a shuffling of AM that essentially provides the same stuff on new frequencies at enormous cost.