Join the debate
We are debating serious questions facing education today..... here are some thoughts to get you going.
Why not come and join us at the 2020 FT WEEKEND OXFORD LITERARY FESTIVAL EDUCATION LEADERS CONFERENCE: ASCENT OR DECLINE? EDUCATING THE WESTERN MIND IN AN AGE OF CHINA, TECH AND AI to carry on the conversation?
Will soft skills be highly employable in an automated age?
Automation, driven by machine learning, will shortly inform 90% of human decision making and judgements. Soft skills will be more like the ‘gloss’ with which we express these decisions informed by data. Yes, soft skills will improve the tone of that communication; yes, they can build rapport and smooth over cultural barriers and increase the pleasure of work- but these are benefits generated by general and fairly low level social cognition. In other words, most people will already have the ability to accomplish them. Like cereal crops, these soft skills will be ubiquitous; required, but not precious. The idea that having better soft skills will make you more employable in an automated is a myth.
Soft skills will continue to drive work productivity: the quality of human interactions drives efficiency of communication; increases motivation; creates belonging; coheres alignment. Most business decisions are made collaboratively; arriving at a shared judgement involves subtle, complex, emergent collective checks and balances. Machines may supply raw data but only with human judgement will bad decisions be avoided, and good decisions be made. Understanding collective emotion and making good judgements will be invaluable skills. In addition, soft skills creates human pleasure: we work together for more than just getting things done.
Will biases in AI destroy society?
For two reasons. First, AI algorithms are trained on datasets. Those data sets are selected by human beings from many, many possible data sets. In that selection bias occurs, partly by having incomplete datasets to choose from; partly because of conscious biases the coder himself has; partly because of unconscious biases the coder has. Second, AI algorithms are self-learning. This means the decisions they arrive at are hidden inside a black box; not even the coder who wrote the algorithm can explain how the machine arrived at the decision. Neither can the machine. AI builds in undetectable biases which will shape every decision from financial mortgage applications to criminal detection. There is no way out.
For two reasons. First, we can make AI algorithms more transparent than they are. There is work going on to simplify the self-learning within machines, such that we can better read-back how they arrived at their conclusions. This approach offers some promise. Second, we should look at other ways to measure bias. For example, imagine that banks were forced to make visible their mortgage data, showing the characteristics of those they offered mortgages to (as a group not as individuals). Using this data, it would be simple to compare this with a wider population and see if the bank was favouring one group over another. Another solution, which STEER is working on, is to capture the emotional impact a news page has on its reader’s minds. Using this data, it would be simple to show if a particular news site was consistently biasing reader’s emotions toward a certain perspective. Such methods can unmask the invisible biasing effects of the algorithms shaping the world we live in.
Can education measure soft skills?
Soft skills are a bucket of behaviours and capabilities which should be seen as virtues recognised by society, not science measured by data. Attempts to define soft skills often prove to be little more than branding exercises: Emotional intelligence proved to hollow because the brain doesn’t process social-emotional cognition in a discreet intelligence as the idea suggested. Skills such as persuasiveness, influencing and apparent empathy traits can also be exhibited by psychopaths. Character has proved elusive; who defines it? Surely character can only be recognised by other people not by oneself? The person who self-evaluates as of noble character is, by definition, no such thing. Values evaporates too - whose values? Hitler’s or Gandhi’s?
The mistake in C20 thinking about measuring soft-skills is to see them as personal attributes. Soft skills are only such when they produce a positive collective social outcome; they produce collective pro-social results. You can measure soft skills by measuring those collective outputs, rather than trying to measure the inputs. Such social outputs might be indicated by increased numbers of relationships; greater intimacy/relationship; higher trust/relationship; increased longevity of relationship; ability to overcome setbacks; increased diversity of ideas in a population.
Will the Chinese curriculum win?
Abundant evidence indicates that, especially in mathematics and hard sciences, Asian students accelerate beyond their Western peers in knowledge and learning. This will give them a competitive advantage for jobs which require precise, qualified, technical expertise. Such expertise cannot be guessed at or approximated. Whilst there may be some disciplines better suited to more creative learning, the high value economies of the future will be driven by those who understand the science.
The dichotomy between science and arts is false. All learning is creative as well as scientific. For example, a mathematician must be creative when looking for new patterns to study; an artist must be mathematical when teaching others how to cultivate, or appreciate, the art itself. The 15th Century renaissance exploded from the interaction between artists, theologians, scientists and mathematicians whose work informed other disciplines. The West must regain its confidence that innovation and discovery comes from integrating arts, sciences and humanities and not segregating them into economic/ uneconomic occupations.