AI predicted the U.S. elections better than humans

Posted Nov 20, 2016 by Tim Sandle
Trump or Clinton? The opinion pollsters generally called the result of the vote wrong. In contrast, an artificial intelligence system called MogIA predicted the result to a high degree of accuracy. Does this mean more investment in AI systems?
The artificial intelligence (AI) system MogIA was developed by an Indian start-up company called The system, as reports, made an accurate production of the U.S. election by trawling through 20 million data points. The information was collected from Google, Twitter and YouTube, and other social media.
To form the prediction the AI platform reviewed public engagement throughout social media with respect to the candidates running to president. The reasoning is that the more a person engages with social media, especially in a positive way such as ‘liking’ a Facebook post or re-tweeting a tweet then the more likely it is they will act in favor of the information. So, whereas a slightly lower number of people said they would support Trump to opinion pollsters than those who backed Clinton, the on-line activity of people suggested a slightly higher weighting towards Trump.
The ‘intelligence’ factor is not only with tallying up the various ‘likes’ but with sensing and interpreting the behaviors of people. Thus MogIA becomes a method for investigating how what people say and think reflects, and perhaps even affects, their voting intentions.
Political polling is not the only area where MogIA has been successful. The system has been able to predict all of the winners at the Kentucky Derby. Drawing on something else entirely – the decision of judged – the AI system has a 79 percent accuracy in predicting the decisions made by the European Court of Human Rights.
Speaking with CNBC, Sanjiv Rai, who founded stated: “While most algorithms suffer from programmers’/developers’ biases, MoglA aims at learning from her environment, developing her own rules at the policy layer and developing expert systems without discarding any data.”
The MogIA system was created in 2004 and it has been getting progressively ‘smarter’ as time has gone on. Perhaps rapid advances in processing big data and machine intelligence could replace the standard way of assessing public opinion in the future.