>Traditional ASR systems struggle when English (or any language) is spoken with a heavy accent, often confusing it with another language.

Humans also have difficulty with heavy accents, no?