# Thinking ahead to a future with more mathematical thinking

To continue the theme started with my review of *“The Model Thinker”*, I challenged our guest bloggers to reflect on more mathematical thinking.

Ever prompt to respond to such an opportunity, Tony Boobier has below shared his reflections on this topic. Readers may recall that Tony is a globally recognised expert on AI & Analytics, especially advising senior executives on how to transform. I’ve previously shared reviews of two of his books on this topic, one focused broadly on the future of work & one specifically on banking.

In his usual musing, quirky questioning style, below Tony asks all leaders to consider why and how more mathematical thinking might be needed. How the changes coming to our businesses mean more than implementing new technology. Rather, how preparing for the greater role of AI also means expanding our approaches to include more mathematical thinking. I hope some of his questions get you reflecting on changes you could make in your workplace.

## Do we really need more mathematical thinking?

In considering the topic of applying mathematical thinking to real-world problems, I wondered if it was as much a philosophical issue rather than data, analytics and ultimately AI.

It implies that mathematical thinking is or at least could be, at the heart of everything we do. Whether we are conscious of it or not. For example, is intuition simply a mathematic process by which our brains weigh up the odds of success or failure? If so, can’t it be reproduced by a computer? Isn’t online dating no more than some sort of digital analysis of the way that humans form relationships?

Some professions, such as insurance underwriters, already use mathematical thinking to consider the risk which might attach to an insurance policy. From this, they set an appropriate rate or premium payment. They would say that their business is entirely based on effective mathematical thinking. Incidentally, my second book “*Advanced Analytics and AI: Impact, Implementation and the Future of Work*“ tabulates the professions most at risk from the Data Revolution. It specifically highlights insurance underwriters as a profession ripe for a high level of automation. In other words, the insurance underwriter might already seem to be on the list of endangered species. Perhaps a new series for David Attenborough looms?

## Does Hitchhikers Guide to the Universe have the answer?

The author Douglas Adams posed the question in Hitchhikers Guide to the Universe, *‘What is the meaning of life, the Universe and Everything?’* Few readers expected that the answer would be a numerical one. That particular answer was generated by a computer which Adams called Deep Thought. One problem was that it took the computer 7.5 million years to come up with the answer. Probably that length of time could nowadays be reduced by the emerging use of Quantum Computers.

A numerical answer to what perhaps is the hardest question of all seems to be a good (but extreme) example of mathematical thinking. Nowadays, more simple questions can be churned out in a lesser time. For example, the quickest way to find our way home and avoid traffic jams.

Deep Thought was also asked to produce the *‘Ultimate Question’,* which it could not. Instead, it said it could build a new super-computer which incorporates human beings into a *‘computational matrix’.* That turns out to be Planet Earth. Of course, all this is literary nonsense, and Adams joyfully plays with his readers.

## Mathematical thinking and the rise of AI

But the idea of using mathematical thinking to solve real-world problems is not so wacky. Already, organisations use mathematical simulations to stress test the banking system. Some would say that the resilience of the banking industry to recent economic stresses is a good case study in using mathematical thinking to keep society ‘*ticking over*’ financially.

In a different scenario, consultants Booz Allen claim to help the US military to transform their wargaming capabilities by using experiences enhanced by artificial intelligence (AI). Their particular aim is to enable rapid experimentation. Through this approach, warfighters are able to make quicker better-informed decisions to defend air, sea, and land.

Turning to today, although the mantra seems to be to use more Artificial Intelligence, many are already thinking it is starting to get over-hyped. Closer to home is the increasing use of data-based Robotic Process Automation (RPA). RPA is simply a way of using mathematically modelled data and analysis to make business decisions and automate outcomes. Experts suggest that RPA will replace humans who are carrying out menial, repetitive tasks. But we shouldn’t overlook that these humans are themselves dealing with, and solving, ‘*real world problems*’.

## Dangers lie ahead if we are not careful in our thinking

Not all is so rosy in this new automation. Scientists continually reinforce the mantra that ‘*the truth is in the data*’. One challenge however is that data can often be biased, either consciously or unconsciously, and thus lead to a skewed result. This distortion can occur in many ways, including how that data is collected. For instance, data is collected from an inadequately representative sample.

Biased data leads to flawed calculations. The problem may not be in the mathematics itself but rather in the source of information that the maths is based upon. Machine Learning systems which themselves are based on biased data may only make the problem worse. We may even find ourselves with mathematically correct modelling, but still giving us the wrong answer to real-world problems.

Where does all this leave us? Firstly, mathematical thinking potentially can replicate human thinking, or at least, supplement it. We are first moving to a world of ‘Augmented Intelligence’ rather than solely artificial intelligence (AI).

## Where will you go on this journey? What are you considering?

Perhaps finally, we should cast an eye to the future. Data growth is currently exponential. Before long we will have an Everest of data to manage, to interrogate and to provide insight. Synthetic data may help us get over the problem of bias. More effective and quicker computing will not only provide greater insights but will also underpin rules-based automation. The use of data-driven mathematical thinking to serve all real-world problems may not be completely with us yet, but it might be closer than we think.

Thanks, Tony for those thoughts. It is a different angle than the topic I had intended. But I can see the benefit of us reflecting on such questions and preparing for a different working future. So, thanks for making us smile & think again.

What about you, dear reader? Have you been reflecting on how you could apply more mathematical thinking to your challenges? Did my suggestions for more realistic behavioural analysis, or Tony’s questions above spark ideas for you? If so, I’d love to hear how and why you are applying more rigorous mathematical or statistical thinking to real-world problems. Perhaps you could be our next guest blogger?