Asymmetrical AM (Amplitude Modulation).

[Mike Watterson]

That would be asymmetric distorted audio.

That's what I thought, too - unless the audio waveform is complex and asymmetric, and contrived so that the positive excursions are higher amplitude than the negative (maybe they're narrower, so that the average is still zero).

What I meant, is in the link. You can't even go quite to zero carrier and if you measure as per the linked article for say 1% or 2% carrier minimum, it's rare audio that is net asymmetric as that would be a DC offset. The AC coupling will remove it.

Indeed you only need to look at a minimum carrier level. Getting to zero is a total fail and bandwidth (splatter) and distortion will jump.

In no case shall the amplitude modulation of the carrier wave exceed 100% on negative peaks of frequent recurrence, or 125% on positive peaks at any time.”

The author of the essay doesn't consider adaptive carrier power, which makes it seem like you get more than 100% if it's done wrong. It's not due to "asymmetrical" peaks on speech. On a pure AM system that never loses carrier there is no asymmetry because there is AC coupling. It doesn't matter what sort of loudness compression you have (they also sell adaptors for CBs, Ham and mobile radio and they always reduce readability because they over compress).
See https://www.eham.net/reviews/view-product?id=2343 W4RT Electronics One BIG Punch (OBP).