Hindsight · review build · data not yet validated
The majority of text is created using continuous speech recognition (CSR) dictation software, but keyboards are still used. CSR is very accurate, far more so than the human transcriptionists who were used up until a few years ago.
Grader A: miss
In 2009 almost all text was typed. Dictation software was a niche tool. Speech recognition reached parity with professional human transcribers on the standard conversational benchmark only in 2016 (5.8% vs 5.9% word error). Voice assistants arrived with Siri in late 2011 and did not make speech the main way to create text.
Test: Majority of text created by dictation: well under 5% of text in 2009 and 2014; CSR accuracy below human transcribers until 2016.
Grader B: miss
Speech dictation (Dragon) stayed a niche tool; keyboards created almost all text in 2009 and in 2014. Machine transcription reached human parity only on a 2017 benchmark.
Test: Share test: text created by dictation well below 5% by 2014.