Billionaire New York businessman Donald Trump might start suddenly become a fan of artificial intelligence developed in India.
The artificial intelligence (AI) system, which has been developed in Mumbai has successfully predicted the last three US presidential elections and this time round, the system has put Donald Trump ahead of Hillary Clinton in the race to the White House.
The system, called MogIA, was developed by Sanjiv Rai, the founder of Indian start-up Genic.ai. Sanjiv took 20 million data points from various platforms including Google, Facebook and Twitter and performed sentiment analysis on the data to create predictions, reported CNBC.
MogIA looked at the engagement of tweets and Facebook Live videos posted onto various platforms and deduced that Trump will not only win the election, but he has also superseded President Barack Obama's engagement numbers during the 2008 election by a margin of 25%.
"While most algorithms suffer from programmers/developer's biases, MoglA aims at learning from her environment, developing her own rules at the policy layer and develop expert systems without discarding any data," Sanjiv Rai, the founder of Indian start-up Genic.ai who developed MogIA, told CNBC.
Granted, people didn't engage on social media in 2008 to the extent that they do today and it can be difficult to classify a tweet as positive or negative. Just because someone engages in a tweet by Donald Trump, it doesn't necessarily guarantee that the person will vote for him. However, the system has been correct in the past to its difficult to ignore.
According to CNBC, Rai wrote: "If Trump loses, it will defy the data trend for the first time in the last 12 years since Internet engagement began in full earnest,". Rai also added that there are many more people on social media that there were during the last three presidential elections.
It is becoming increasingly popular to conduct sentiment analysis on social media because of the shear amount of data that is publicly available. In September, Nick Beauchamp published a research paper about an experiment that he carried out that analysed more than 100 million tweets during the 2012 election. He found that the results were very similar to the results from that actual state level polls.
I think its amazing that its even possible to do this kind of analysis. Ten years ago, this would not have been possible. But now thanks to major advances in Artificial Intelligence, cheap and powerful computing power in the form of GPUs and CPUs, and also the fact that the major social networking platforms have released APIs to collect social data, it is now possible. This is very powerful. I'm very exited about this space and as we continue to make progress with AI and Deep Learning, I expect to see more and more of these incredible predictions.
Photo credit: Krassotkin (derivative), Gage Skidmore (Donald Trump), Gage Skidmore (Hillary Clinton), Donald Trump and Hillary Clinton during United States presidential election 2016, CC BY-SA 3.0