Andrew Diddel Blog 1, 9/22/10
IS 251, Tallon
While Google has been flourishing in the online search industry for many years, they seem to constantly be moving forward with new ideas and ways to improve both themselves, and the search experiences of their users. Most recently, as pointed out in Olga Kharif’s article, Google Uses AI to Make Search Smarter, projects like “Google Instant”, “Google Goggles” as well as Google Android’s “Conversation Mode” are all using Artificial Intelligence to simplify online searching.
The piece that I was most intrigued by was the information and discussion about Google’s use of speech recognition technology. Google has certainly shown its power and dominance in the world of online, typed searches, but they are still trying to do more and more to accommodate the preferences of their users. Kharif claims that “the company’s smart software has billions of pieces of data to sift through and learn from.” And with the voice-analyzing software, specifications and refinement of searches can happen based on the consumers “behavior, wants and needs to provide them with more relevant answers.”[i] After reading this argument, I am a bit skeptical of the software’s ability to simplify searches when there is so much going on.
I agree with the idea of computers and Smartphones assisting the consumer by providing more applicable search results, but I am worried about the likelihood of this happening, considering that the smart software already has “billions of pieces of data to sift through”. By adding a constant rush of information that would need to be processed by voice-recognition software, one might think that Olga Kharif may have over-exaggerated the simplicity of completing this task. In David F. Gallagher’s article, Pushing the Limits of Google’s Speech Recognition, readers can see that Google has previously attempted to successfully install “‘the only fully automated voice-mail transcription on the market’” with the “Google Voice” program. The article goes on to show the extensive problems that the speech-recognition algorithms technology in their attempts to transcribe verbal messages. When the technology attempts to transcribe a verbal message like “laying its foundations on such principles”, and accidentally comes up with “sunday schnitzels cripples”[ii], eyes will be opened and questions will be asked.
Questions will also be asked about the company’s new “Google Instant” project. Kharif also includes that according to Google, with the use of AI, “Instant” will decrease the time of an average search by “two to five seconds”. As an interested reader, I found myself very confused when they also said that the time saved would cause people to more commonly use Google. While the company foresees that these extra searchers will increase revenue, I foresee that this idea has more or a downside than an upside. Especially from what was seen with the audio-recognition issues associated with “Google Voice”, cutting down a few seconds here or there might not be worth an incorrect specification or redirection of your search. If I were hoping to research Smartphones, and “Google Instant” automatically saved me a few seconds, but transferred me to sites on the more popularly searched Smart Car, I could be a bit frustrated. What is the point of assuming the direction of a search to save a few seconds, when so many people have constant Internet access with the Blackberry or iPhone in their pocket? Couldn’t they sacrifice a few seconds for an accurate search? I have always been impressed by the continuous progress that Google makes as a company, but I think Olga Kharif might have understated the difficulty of putting these projects into action in the upcoming months.
No comments:
Post a Comment