Today we live in an "information society"
In a bureaucratic society, much communication is textual
Even with advanced recording and playback technology, written reports are more efficient
Human vision is amazingly full of content:
"A picture is worth a thousand words."
Semantic metadata can be included: table of contents, indicies, HTML tags, database schema, etc.
Voice input, yes, but storage will be coded as text for the reasons above, plus efficient search