Artificial Intelligence and Machine Learning: Creator Beware Artificial Intelligence and Machine Learning: Creator Beware
As with anything, these cutting-edge technologies are both fascinating and chilling depending on how they are applied. Artificial Intelligence and Machine Learning: Creator Beware

It feels like the past year in technology has been dominated by talk of Artificial Intelligence and Machine Learning. Just writing those words brings chilling thoughts of SkyNet (the hostile sentient AI network in the Terminator movies) and Minority Report (predictive crime fighting, stopping crimes before they happen and arresting you for something you haven’t done yet) to mind.

But are those enough to make you switch off your computer and drop your smartphone in a glass of water? I’d say no, chilling thoughts or not, because I firmly believe the perceived downsides of these applications are massively outweighed by the value we can derive from them.

Intensive Analysis

Really what these technologies are doing is taking the huge amount of data being created daily and, through intensive analysis, presenting their findings to decision-makers who then use it to guide their organisations’ paths. According to IBM, we create 2.5 quintillion bytes of data in the world every single day. I am not even sure how to put that in my calculator to visualise it, let alone know what do with it.

The reality is that computers and the way that they calculate and interrogate information is inherently better at dealing with large amounts of data and deriving some kind of value out of it than the human brain is. That where Machine Learning and Artificial Intelligence come into play, as computers have far more detailed and efficient ways of interpreting that data. Humans rely on visual information as well as history or learned knowledge to do something with data sets or information. A computer is not biased, and will determine causality in ways only the most intelligent data scientists can explain.

In a recent session with a company specialising in Machine Learning, this was highlighted to the extreme. This company did work for an unsecured lender who was looking to stay ahead of their competition. They had a leading bad debt ratio compared to their competitors, and this was largely due to the experience of the head of loans who brought a lifetime of knowledge with him to help refine the criteria for granting loans. So to extend that lead, a model was built using historical data and analysed by a Machine Learning algorithm; the result was nothing short of astounding.

No-loan Tuesdays

For some reason, one of the recommendations that came out was that people who applied for a loan on a Tuesday were more likely to default on their payments after 6 months and that by denying credit to anyone that applied on a Tuesday, the company would see a significant increase in profits due to smaller debt write-offs. The concept was applied to historical data and it was true!

Now the fact that no one could unravel exactly why that was the result is another story all together. In financial lending, the day an application was made was never thought of as a criterion to determine if you should be given a loan or not, whereas a computer found causality in elements that we hadn’t even considered.

Now to bring it back to a more sinister example. When you put together smart people with lots of data, do you have the ability to change the outcome of an election? In a recent article on Motherboard, it is asserted that psychological modelling created using users’ Facebook likes and activity was used by a company called Cambridge Analytica to craft tailored messaging for President Donald Trump during his campaign. These tailored messages were aimed at undecided voters, and because they carried a narrative that specifically appealed to ideas researchers knew would resonate with them, it convinced many voters to either vote for Trump, or not at all. And we all know how that turned out. Similar tactics were used in the Brexit campaigning.

Polar Opposites

These two examples show the complete polar opposites of what Machine Learning and AI can be used for. One, where the machine is purely fed data and an outcome is produced that can be used to help a business. The other takes the same concept, plugs into our personal lives and marries that data with some smart people and they use it to change the world.

Either way, these results are real and you are being affected by these technologies and concepts on a daily basis, whether you know it or not. It’s time to start thinking how you could be using them in your business and how you can engage customers on the use of technology to fundamentally change the course of their businesses.

Jonathan Kropf is the CEO of Tarsus on Demand.

No comments so far.

Be first to leave comment below.

Leave a Reply