On the backend, Google has taken the audio tracks of YouTube videos of politicians, and automatically converted it to text for search capability.
How will this impact the overall internet? Well, here-to-fore, videos were mostly searched on metadata — that is, text associated with a video indicated what the video was about. Videos posted on web pages might have other, typical webpage attributes that are already used in keyword search, such as Title, H1, body text content, anchor text of links pointing to the page, etc. Now, if/when Google rolls this functionality in a wider fashion to apply to more than just politicians’ videos, any spoken words within videos will add to the keyword content associated with them, becoming searchable.
As you may be aware, Google rolled out “Universal Search” in the past year, their branded version of “Blended Search “. Blended Search in search engines allows keyword search results pages to provide result listings from a multitude of other, specialized “vertical” search information sources. Video search, image search, news search, local search, and other vertical search results are all showing up in search engine results pages these days, whereas those results pages used to display only listings of general web pages.
Considering all the rich verbal content there is to be found in videos, I think we can expect to see video results appearing more and more in the blended search results as time goes by.
Even more interesting/compelling: will Google productize their software for converting spoken content into text? Similar to the Google Translate tool, this could be really useful to many people.