Artificial Intelligence (AI) is high on the priority list of major enterprises across the world. With the famed futurist, Ray Kurwzeil joining Google last month, the buzz on where artificial intelligence will take us in the near future is everywhere. The focus is on machines replacing humans to deliver better and faster results. Scientists across the world have been working on ways to mimic the human brain in computers. However, when it comes to translating human thought into something that the machines can interpret and act on, experts have always fallen short. This is looked up on as the biggest drawback in machines and something that stops artificial intelligence from taking the next step. Machines might go ahead and decrypt human thought, but how will they interpret the emotion behind the thought?
Twitter has revolutionized the way people share news and interact. By offering ads and sponsored tweets, Twitter faced a challenge- How to instantly identify the tone/meaning of the tweets and deliver them to the right streams? The technical team at Twitter faced the same problem that all artificial intelligence experts experience. Normally, topics trend on Twitter for a short period. The machines can identify a spike in a particular query. How will the system understand what a particular hashtag means and identify the type of ads to place against the tag?
Twitter recently said that the machines don’t interpret the meaning of the hashtag. Twitter, instead turns to real people to understand the meaning. Of course, the process is automated, wherein the system identifies a sudden spike and sends the query for human interpretation. Once the meaning of the hashtag is identified, it is then fed into the backend system which will then automatically decide the ads and streams that will suit that particular hashtag.
Based on the current state of artificial intelligence, technology experts should take cue from this story. Rather than trying to replicate human brain, let the machines do what they do best. The best use of analytics and high performance computing is to churn out the data and break it into something that can be assessed and acted upon by a person. When it comes to complex data, analytics and computing is more about automation than offering insights.
Enterprises should recognize what the human beings and machines do best and let them do that. The practical way to look at artificial intelligences is to have machines and human beings work in tandem, rather than trying to mimic each other.