Par Beat Richert
Following on my recent post where I argued that Artificial Intelligence (AI) is much more “electronic empathy” than actual intelligence, hereafter a reflection on the future of education and AI.
During the recent three-day workshop at McGill University during which close to 100 participants imagined and defined the role of the university in 2050. No easy task even for computer scientists, astronauts, biologists and PhD students, all of whom among the multidisciplinary and international group of thinkers attending the event.
One central question (or maybe THE central question) during this inspiring workshop was the human-machine relationship. With robots having been part of our working environment for more than 20 years, there was not much of a discussion about [robots] physical superiority to humans when it comes to precision, reliability and cost efficiency. Of course this question gets trickier when we move to the cognitive capabilities and emotional intelligence of humans. And this is where opinions among participants scattered, dispersing from fascinating opportunity to ultimate thread for humanity.
Admittedly, already the word artificial intelligence (AI) can have a threatening after taste. Most of us are comfortably confident about our master and servant relationship between man and machine. After all, we give orders to computers and they execute them. But of course, this is far too easy. The omnipresence of algorithms in our daily lives has already profoundly undermined our “competitive advantage” to computers when it comes to everyday decisions and intellectual processes. When I talk about algorithms to students, I like to tell that algorithms are no more (and no less!) than pattern recognition tools. While it takes us 15 minutes to empty a drawer full of single socks and match them up to pairs, it takes the computer 15 milliseconds to do the same job. After all, it’s simply about recognizing a pattern [look for two identical socks and match them when found]. When we realize that our own senses are inherently configured and trained as pattern recognition tools (if it has a trunk and if it has branches and if has leaves, it is probably a tree) we start to understand the similarity between computer algorithms and human decision making processes. From Kindergarten (order squares with squares, circles with circles…) through primary school (find next figure: 3,6,9,12,?) to adulthood (IQ test question: If NEW YORK can be encrypted as PGYAQTM, how can you code the word CHARLOTTE?), pattern recognition skills are trained and are an integral part of our education system. Amongst all mental abilities, pattern recognition is said to have the highest correlation with the general intelligence factor of humans.
While you read this sentence, high-frequency trading systems (also known as algorithmic trading) have completed roughly 15 million transactions worth some 50 billion US-dollars. During the same 5 seconds, more than 250,000 Google queries have been made. If we consider that every single question from us to Google is in fact first and foremost an answer to Google from us, it gets obvious that we are teaching the machines to “think”. It is safe to assume that every Google query is analyzed by Google’s machine learning algorithms to constantly refine the relevance of search results and hence fostering Google’s artificial intelligence. What in simplified terms are numerous layers of sophisticated and interdependent algorithms is then perceived as artificial intelligence.
If we now add the fact that much of our daily lives are taken up by habits (approximately 40%) that we’ve formed over our lifetime and if we agree that habits can also be characterized as automatic behavioural patterns, one can imagine the truly unlimited possibilities of pattern recognizing machines. The future for artificial intelligence, based on gargantuan amounts of data and predictive analysis, is incredibly bright! Expect the algorithmic world to be smart, very smart!
So where does that leave us humans? What is the value of information and of our knowledge? What is the future of higher education?
In an attempt to answer this question, I assume that everything that can be digital will be. By converting colossal amounts of information into terabytes of structured data, we open the gates for algorithms to take over our daily search and interpretation of information. Other algorithms then contextualize and customize this information allowing computers to actually “understand” the information, hence resulting in some kind of “knowledge”. Further algorithmic refining (historical comparison of similar personal needs for any given information, comparison with “similar” individuals showing analogical behavioural patterns, use of my geographical position, day and time, use of other factors like people near me, weather conditions, predictive analysis etc.) assures the highest possible relevance and value of the information. Because of its individual and contextual value, the information is perceived as “artificial intelligence”.
The tremendous power of real-time data analytics and predictive analytics upon which AI is based let me conclude that the only cognitive human spectrum that will not be infiltrated with AI will be wisdom. Knowing what we don’t know, intuition based thinking, guts feeling decision. One of the crucial responsibilities of future education will be teaching a holistic understanding of the foundation and the possibilities of digital technology. The efforts of teaching Computational thinking skills are a step in the right direction. Being aware that the digital world we are living in is a model of our real world seems to me a fundamental part of wisdom. Models are by definition always a simplified representation of the original, hence leaving a wide open space for wisdom.