Transcription combines speakers

I've been using the transcription feature to create the captions for my YouTube channel, and it's largely a lot of help. But the speaker detection tends to merge speakers together, especially female speakers. In my footage, there are 3 female speakers: 2 with Australian accents, and 1 with an American accent. But Premiere Pro doesn't distinguish between the three of them, even though one has a completely different accent!

Repro steps:

1. Record a video with multiple speakers, both male and female.

2. Use Premiere Pro's audio transcription feature to generate captions.

3. Notice that similar sounding speakers get treated as the same speaker. The effect is more pronounced for female voices than male voices. (Perhaps it is pitch-related?)