A conference organizer wants to display live on-screen captions in the same language for hearing-impaired attendees while a speaker talks at the podium. Which AI capability should be used to convert the speaker's spoken words into the displayed text?