Atlas record: https://www.aitimeline.dev/en/events/evt-0212/
Availability · November 2008
Google Voice Search
Google released an application that sent recorded speech to a server and returned search results, with no training on the individual user.
Why it matters
Speech recognition became free and ubiquitous, and its quality now depended on how many recordings had been collected.
The acoustic model was assembled over three years from the GOOG-411 directory service, which Google ran precisely for the voice samples. The computation happened on the server, so the phone could be weak. That closed a loop: more users give more data, more data gives a better model. Three years later acoustic models began to be replaced by deep networks and quality jumped again.
Event record
- Event date
- November 2008
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 17, 2026
- Lines
- ID
- evt-0212
The application was released in November 2008.
Related events
Earlier
- Related Dragon NaturallySpeaking 1997
Recognition moved from the personal computer to the cloud, and per-speaker training became unnecessary.
- Builds on The Switchboard corpus 1992
The model was trained on spontaneous telephone speech rather than read text.
Later
- Related Siri 2010links to this record
From a spoken query for search to a spoken query that performs an action.
- Related Siri in the iPhone 2011links to this record
A second system making voice an ordinary way to address a device.
- Related Amazon Echo and Alexa 2014links to this record
Recognition happens in the cloud, as in voice search.