I would class live transcribing and captioning as the same thing.
I can see a user need in here which is the ability to carry AAC symbol sets for deaf people with learning disabilities (or otherwise complex needs). A user may require a specific (and sometimes proprietary) set of symbols so any RTC system would need to be able to load a symbol set (you couldn't rely on pre-loaded ones).

Here is the suggested new draft text: • REQ 12a: Honor user preferences relating to live transcription, captioning and provide support for signing or type of symbol set e.g. AAC.

Suggestion to go back to the OP

Discussed on RQTF call https://www.w3.org/2021/01/20-rqtf-minutes.html

Reply from ITU-T:

The text is okay, however AAC is not a symbol known for live captioning, it is usually CC or lately [=]

@RealJoshue108 notes that which AAC may not be used for live captioning it may be used for captioning in general. So we wish to include it as a user preference that needs to be supported by user agents.

@RealJoshue108 I think we need to wordsmith this a little further, such as:

"REQ 12a: Honor user preferences relating to live transcription, captioning, and provide support for signing or use of symbol sets e.g. AAC."

Reading the above, I do wonder if we need to separate out the requirements for live vs prerecorded content? This thread does seem to indicate subtle but important distinctions.

The distinction in WCAG 2 is between captions for live and captions for prerecorded content. The term "captions" is used in both cases. This is why I suggested in the meeting yesterday that we should align our terminology with WCAG. It seems to me that distinguishing "captions" from "live transcription" here is likely to be confusing to anyone familiar with WCAG. Also, captions may be verbatim transcripts or may be edited, and we should allow for both - especially in a live event such as a WebRTC-based meeting.

Taking a steer from @jasonjgw I'm going to suggest we add this to the draft. It removes the term 'Live Transcription' in the heading and refers to 'Captioning support' - then separates out the different kind in the suggested requirements. I think this reads better and should be less confusing. Thoughts?

Captioning Support
User Need 15: A deaf user or user with a cognitive disability needs to access a channel containing live transcriptions during a conference call or broadcast.
REQ 15a: Honor user preferences relating to captioned content also provide support for signing or use of symbol sets e.g. AAC.

I would start a new sentence after "captioned content".

Done. Thanks @jasonjgw