The Answer Engine: How Humans Can Provide Better Knowledge Than Algorithms
Editor’s note: Michał Borkowski is CEO of Brainly, a social learning website.
When you have a question you need answered, do you ask a friend or a robot? The answer to that question used to be easy. But the rise of Siri, Yahoo and Google has ushered in an era in which we turn not to experts but to search engines for answers.
By failing to solicit human responses in the pursuit of information, I believe we are making a big mistake – in relying excessively on this single source, we open ourselves up to a data blind spot.
When the first hand-curated directories of the early web’s few sites added query forms, becoming search engines, they expected users to enter boolean search strings, just like their computer scientist inventors. But the fact that many of us write questions, not fact-focused strings, into search boxes is a dead giveaway. Humans want to use natural language when they solicit help.
“How tall is President Obama?” Through algorithm updates like “Hummingbird,” Google has evolved to recognize such questions and, increasingly, to answer them itself.
However, for an algorithm to answer questions like these it must first have access to plentiful underlying data, structured in a way that it can be queried. This is not the case for the vast majority of search engines and won’t be for some time if ever. Ultimately, meaning most humanized question queries will lead users to static web pages.
Humans want to use natural language when they solicit help.
Now is the time to bring humans back into the web. At a time when the new Internet of Things trend stands to bring many more machines on to the network, we should realize the valuable role we, as operators, can play in our own creation, asking each other for information and supplying it in the spirit of cooperation.
The problem with robotic search engines is twofold. First, search result signposts have inaccurate understanding of timeliness. The algorithms’ trainers have worked hard to focus their attention on fresher web pages. This can down-weight “stale” information, but it can also falsely present very recent results simply on the basis of freshness, overlooking a five-year-old web page that was more relevant.
Second is the depth issue. When I first started using the Internet in the 1990s, I participated in many special-interest forums. It was great to find like-minded users with deep knowledge in niche areas, seeking help and offering support by supplying detailed responses. Sadly, this seems to be a lost era. Today, search engines and networks operate on a web-wide basis, with little understanding of niche knowledge to match the expertise of specialist practitioners.
When you aggregate people, not just web pages, you unleash something wonderful. You allow them to be each other’s search engines and open up deep specialist knowledge and community spirit.
Creating these conditions on today’s web can be tricky. I think many people ask questions of Google, privately, because they are embarrassed to display their lack of knowledge by posing a question on a web they know is all too public.
What’s crucial for creating people-powered human answer engines is the community of users.
Fortunately, several services are overcoming this factor. Yahoo Answers may have earned a poor reputation for the quality of some of its community’s responses. But creating an environment in which no questioner or answerer is discouraged is vital to creating a web in which human peers take on the task of information delivery. Allowing even the simplest question to be posted in a positive environment is what Yahoo has cracked, creating a mechanism that makes users comfortable with their own contributions. The stats don’t lie – Yahoo Answers’ traffic keeps growing, currently at more than 170 million monthly unique users in the U.S.
Quora, a question-and-answer site, has cracked the nut even harder with resident experts and clearly defined guidelines for contributors. Its growth reflects its popularity. In April 2014 it raised an extra $80 million and is valued at nearly $900 million. Traffic to the site in February 2014 neared 2.2 million visitors solely from the U.S. Quora’s global success can be attributed to several factors, but much of its focus is on our culture’s insatiable desire for the truth.
What’s crucial for creating people-powered human answer engines is the community of users. In Brainly’s user base for example, we know that everybody, no matter what their age, has skill and knowledge that can be shared. This is the substance of crowd learning. The need for truth and being part of a learning community is growing, as we’re reconnecting with people, albeit through technology to reach this goal.
It’s refreshing to know that in this seemingly open-ended world of search engines, algorithms and unregulated forums, Q&A platforms with reliable and knowledgeable contributors still have a very large part to play.