Abstract:

Creativity has been a focal point for educational psychology for over 70 years. Initially framed as a counterpoint to narrow, convergent, notions of intelligence – in other words, human intelligence is more than just the ability to produce correct answers to closed questions – creativity has been promoted as a necessary adjunct to the traditional focus of education systems. However, it is only in the last decade, since the advent of digitalisation and the 4th Industrial Revolution, that creativity has enjoyed a fresh, and more urgent, focus. With the emergence of Artificial Intelligence (AI) that is finally beginning to deliver on its claimed potential, creativity is now recognised as an essential 21st century skill. The reason it is so essential is that creativity is (notwithstanding some claims about AI) the last bastion of human mental sovereignty. It is one of a small number of abilities that humans possess which are essential to the future of work, and which cannot be automated.

The fresh enthusiasm for creativity in education, however, brings some fresh challenges. If creativity is to form a much stronger focus in education in the future, then schools and universities must have the means to assess it. Not merely in theory, but in practice. Unfortunately, current assessments of individual creativity fall well short of what is required in practical, educational settings. The problem is not that individual creativity cannot be assessed, but that creativity assessments are not fit for purpose.

This paper will discuss five factors that impact on the fitness-for-purpose of individual creativity assessments for educational applications. These include:
(a) domain-specificity – the need for creativity assessments across different parts of the curriculum;
(b) consistency and trustworthiness – the need for creativity assessments that do not depend on inconsistent human raters;
(c) authenticity – the need for creativity assessments that integrate smoothly into crowded curricula;
(d) speed – the need for creativity assessments that provide rapid results, paving the way for use as formative feedback, and
(e) cost – the need for creativity assessments that do not impose insurmountable barriers to their use in schools.

The paper will culminate by discussing the development of two automatic assessments of individual creativity that address the shortcomings listed above. The first, an image classification approach developed to assess figural creativity, and the second, a natural language processing approach to assess mathematical creativity, will be explained in detail. The results of these approaches with be discussed, demonstrating that AI-based approaches to creativity assessment can outperform human assessments, with much greater speed and lower cost. These tools enable schools and universities to focus more effectively on the development of creativity as a vital 21st century ability.

Keywords:

Creativity, Assessment, Skills, Future or Work, AI, Digitalisation.