When less than all participants of a remote conference consent to being recording, meeting minutes may need to be manually recorded as text for those non-consenting participants to ensure a complete record of the meeting. By signaling a minutes taker via their input device that a current speaker is a non-consenting speaker, the input device may enable or emphasize a textual input component and/or indicia of the current speaker. As a result, a conference server may receive textual content in place of speech for non-consenting participants and record the conference that is complete but omits speech provided by non-consenting participants.