The Blog

Sit Up And Listen, AI Predicted Trump And It's Predicting Other Things Too!

AI is an exciting world right now. However, like so many things in tech it is misunderstood by consumers. When you say Artificial Intelligence a lot of people think of a robot that talks to you, but actually it's used almost constantly in daily life already.

I sit here writing this aghast at the results of this week. I, like many, was worried about the 'what if's' around Trump. That being said, I was equally concerned about Clinton as well. Had I had the opportunity to vote, is the lesser of two evils really a vote at all?

The Millennials vote, Sanders, was long out of the game and, even with throwing his support behind Clinton, failed to sway the masses with over a third of the electorate in this category not really turning out at all, rather, they like others abstained from voting.

We are in a strange world right now, no longer are people comfortable with the status quo, and the democracies around the world are turning to fierce polarised leaders to 'reset' their countries and make a change.

And you know the crazy thing? Artificial Intelligence (AI), machines and software with no agenda, allegiance, heart or soul predicted it all... and we didn't listen!

OK, let's back up a little bit!

AI is an exciting world right now. However, like so many things in tech it is misunderstood by consumers. When you say Artificial Intelligence a lot of people think of a robot that talks to you, but actually it's used almost constantly in daily life already. It is the process of any software learning and then creating outcomes based on this learning. The ads on Google and Facebook directed at you are AI as too are many video games when the enemies learn your tactics in order to out manoeuvre you. In fact, every day you are interacting with thousands of machines that are learning your every move, everything from what route you take on the Subway/Underground to the the groceries you purchase FreshDirect/Ocado. Make no mistake big brother is all around.

The fear that professors such as Stephen Hawkins have is that AI at some stage will become so intelligent that it re-programs itself, other systems around it and eventually starts to reproduce. Seeing humanity as an imperfection it will eventually, quickly and efficiently wipe us out in a Terminator/Matrix style way. They are likely right, which is a scary thought, and should this happen the process will be exponentially fast. That being said the reality is in the mean time we back in the less apocalyptic world need to get on with our lives!

The reason I believe Professor Hawkins' theory (albeit not right now and hopefully outside my lifetime) is correct is down to our own stupidity and a total inability to think impartially on a subject.

It's been well documented over the past few days that MogIA and AI system developed by Sanjiv Rai, correctly predicted not only that Trump would win, but also the fact he would win the primaries beforehand. In fact, the system was so confident it (and Sanjiv) called it back in October.

'October?' I hear you gasp, yep, the system Sanjiv developed has for the last 12 years correctly predicted all four elections in its lifespan. So a system smart enough to predict the last three election winners based on data it scraped from the internet unsurprisingly predicted the fourth as well and yet we didn't take notice?

Come on human race we need to take note!

Perhaps it was because we weren't impartial, or perhaps it was because we didn't trust the software, but in either case it was right and we were wrong. This is an important point to note as this will be the eventual human demise as AI learns see these losses to us as imperfections.

The data for these marvellous predictions comes from the internet and MogIA looks at over 20 million data points from an array of sources like Google, Facebook, Twitter, and YouTube.

So do we think if it had more structured data sets, something a little more grounded that perhaps we might sit up and listen to it? Where might data like that be?

Well you don't really need to look much further than the stock market for it too has an impressive 86% chance of predicting who is gonna win.

Michael Snyder wrote a blog post back in August that pointed out 'if Standard & Poor's 500 stocks go up during the last three months before an election, the incumbent party almost always keeps the White House'. This metric has been right for 19 of the last 22 elections with only three elections being wrong.

Needless to really say here but the last three months have not been good for the S&P 500 and again predicted the incumbent party would not keep the house.

So if MogIA used more, consistently strong data sets like this would we have taken more notice of this brilliant piece of engineering and take its advice?

I dare say pig headiness, doggedness and determination to drive our own agenda probably would have got in the way.

So with AI becoming such a strong feature of our daily lives what else will it be predicting for us in the future? Well, honestly virtually anything, my phone knows before I do which app I want to use, my fast food app already knows when I want pizza, and my taxi already knows where I want to get to. What I am fearful of is the stuff it is already planning for me that I don't know ..... eeek!

Before You Go