Ever search for a phrase or a definition only to find Google offer up the answer in a handy snippet of information right below your query but in front of the first search result? Its an algorithmic process thats designed to use Googles massive database to causally dip into a wide variety of tasks. The featured snippets tool is designed to quietly replace one of the internets biggest sites, Wikipedia. But its failing.
One reason Google works so well is that it offers entire pages of results, letting people pick out the right one. But the featured snippet is just one result, and backed up by Googles ubiquity as the all-seeing eye of the internet, it appears foolproof. It also offers up some truly terrible and incredibly inaccurate results.
The Outline reported on the problem earlier this year, and Google has since removed some of the worst snippets named in the article. They used to offer up five U.S. presidents in response to a google search for any presidents in the KKK, despite the fact that none were. A snippet also once affirmed that Obama was planning martial law, another falsehood pulled from a disreputable site. Still up? One snippet claims that MSG is dangerous in response to a search for MSG dangers, despite the fact that the stigma can be traced back to a singlescientist who ate Chinese food once, got sick and hypothesized that the additive was responsible.
The reality is that, while Google can manually remove the featured snippets that gain notoriety, the process is automated featured snippets pop up for an estimated 20 percent of the millions of searches users can make on the search engine. Wikipedia may not be an entirely trustworthy source, as any middle school English teacher is fond of pointing out, but Google is nowhere close to replacing its network of human volunteers.
Writing in a forward-looking end-of-the-year article from a collection of Microsoft researchers, SusanDumais, a scientist and deputy managing director at a Redmond, Washington, research lab, made a compelling prediction for the future of search engines in the next decade.
The search box will disappear, Dumais said. It will be replaced by search functionality that is more ubiquitous, embedded and contextually sensitive. We are seeing the beginnings of this transformation with spoken queries, especially in mobile and smart home settings. This trend will accelerate with the ability to issue queries consisting of sound, images or video, and with the use of context to proactively retrieve information related to the current location, content, entities or activities without explicit queries.
As we head towards this prediction perhaps even sooner than 2027, Googles featured snippets problem has yet to be addressed. It could get even tougher to root out if a similar approach takes over voice search as well.
The search engines data collection has allowed it to casuallyswallow up entire business models. The most recent example: Google for Jobs, which aggregates job positions from a vast selection of different job boards. Granted, the job boards still serve as intermediaries, since Google just redirects users to each individual job posting, but the job boards lose out on large amounts of data on how many users are searching, and for which positions. If Google wants to create its own job board, it already has more data than its competitors, and it can block them from gaining more.
Read our reviews on Google Home at TechCo
In other cases, Google can destroy entire businesses. CelebrityNetWorth.com is one example: the business calculates how much money celebrities are worth. Or it did, as a follow-up Outline article covered, before Googles featured snippets started delivering the information to searchers, denying page views from the sites that actually did the math.
But the limits of Googles steamrolling algorithm can be seen in every wrong answer they turn up, and this fact is never highlighted better than in its featured snippets tool.
Read more about whats happening at Google at TechCo
See the article here:
Google Is Trying to Be Wikipedia, and They're Failing - Tech.Co