The obvious answer is of course yes, as in philosophy we are already struggling for 2500 years to address some of the most critical questions of life. Why is there life, what is it purpose, are we alone in the universe, what is the purpose of the universe, is there perhaps a multiverse and so on. Of course, if we indeed would be able to ‘know all’, we wouldn’t be humans but some sort gods and knowledge management would not be necessary.
However, what triggered me this time was an article written by my much-respected American colleague David Weinberger in Aeon. Here he talks about machine learning and that brings us back to the relationship between technology and knowledge.
This takes yet another turn from the article I wrote earlier: How humans employ technologies and data to work, govern, and co-exist.
What we are learning from Machine Learning is that if we expose computer programs to an enormous amount of variable data, in an unstructured way they are provide outcomes that we as humans are often unable to understand, yet the outcomes provided are valid. ‘Normal’ computer-based systems are based on rules written into the software programming. In Machine Learning it are not people, but the software programs that are writing the programs.
David explains it along these lines: The software developers provide the computer with thousands of bits of labelled information. The system than algorithmically discovers statistical relationships. The more data the more accurate the relationships and therefore the outcomes.
It is simply impossible for humans to handle such an amount of data in any comprehensible way, so again we are running against our own knowledge wall.
I was also thinking about the chaos theory here, with the classic example that a butterfly flapping its wings here can potentially (under all the right. totally unpredictable circumstances) create a storm in Chile.
We all experience that a relatively small and local weather event, a car accident or a train derailment can have far more reaching consequences than the actual incident itself. It is impossible to know what all the (unintended) consequences will be and then again even the slightest change in the accident pattern will again have totally different outcomes.
We instinctively know that the complexity of life on earth, the universe and all its variations do have an enormous influence on us, but we take this for granted and (fortunately) don’t think too much about it. However, if we stand still, we realise that we know much less than that we think we know.
We are also learning from the field of quantum mechanics that there are stunning patterns apparently developing out of what looks like chaotic developments, from the formation of galaxies to the veins in the leaves of trees. Again, we don’t understand this but quantum mechanics is another key tool that we need to be able to read better the world around us.
Yes, we have created a range of rules-based scientific laws and there are great mathematical equations that describe the various natural phenomena. Machine Learning shows that if we use a wide range of data from the most diverse sources, we can get better outcomes for example in managing our environment. True sometimes these big data events do create unintended consequences and show biases but in general they are rapidly becoming new tools that we desperately need to for example address climate change, pandemics, cancer treatments and so on. I would argue that without big data and Machine Learning we will be unable to address these issues in any manageable way. And of course, it is not a surprise that machine learning is already heavily used in these sectors.
On a day-to-day level, we are of course experiencing the outcome of machine learning when using our apps on our phones, the social media are totally driven by machine learning and their advertising revenue model totally depend on this. The fact that these companies are now among the richest in the world shows that at least from a commercial perspective machine learning is bringing in the dollars.
Another sector that totally depends on machine learning is the financial market, algorithms in fractions of seconds are making mindboggling decision, without any human intervention.
But the good news often gets less attention, in the recent pandemic two Covid Vaccines were totally developed based on machine learning. As a matter of fact, machine learning is one of the key reasons why our medical scientists were able to develop vaccines in a fifth of the time that would normally be required.
Of course, it is understandable that we are uncomfortable and often worried about processes that we can’t comprehend ourselves and must rely on outcomes developed by computers. It is part of the wider phenomenon of post-modernism (alternative facts, lack of trust, ant-rationalism, discomfort with reason). But here we come to the crunch of the issue, are we able to except that there is a limit to our knowledge and that we do need machine learning and big data applications to obtain a better understand of the world around us?
At the same time, we need a new narrative for the non-material elements of us as human beings. Religion, populism, autocracy is all tapping into this discomfort, but what is needed is a much stronger and much more globally accepted worldview of where we human beings fit in all of this, both ‘materially’ and ‘spiritually’.
In our world where complexity increases by the day, knowledge as such becomes less relevant. The only way we will be able to address our hyper-turbulent environments is to create far greater interoperability between people, systems, environments, New technologies such as machine learning, big data, AI and quantum mechanics are the tools that will enable us to take us to the next level beyond pure human knowledge.
Paul Budde