Artificial Creative Intelligence (ACI)

In 2005 and 2011, educational psychologist Kyung Hee Kim published two notable articles for those of us currently considering Artificial Creative Intelligence (Chen et al. 2020). In the first, she said:

The negligible relationship between creativity and IQ scores indicates that even students with low IQ scores can be creative. Therefore, teachers should be aware of characteristics of creative students—this will enable teachers to see the potential of each child. (65)

In the second, she said:

… creative thinking is declining over time among Americans of all ages, especially in kindergarten through third grade. The decline is steady and persistent, from 1990 to present (293)

In reading Kim’s findings, we are presented with the intriguing possibility that our decline in creative thinking has some link, somehow, with contemporary digital technologies, because the emergence of these technologies correlates with her detected fall in human creative thinking – that is the emergence, specifically, of technologies that are now fueling the creation of contemporary artificial intelligence (AI).

The evolution of AI (first named in the 1950s, but evolving since, into our contemporary iterations of AI), combined with thoughts on Kim’s findings, additionally might bring us to wonder whether we have long mislabeled ‘intelligence’; or, perhaps more accurately, whether we have labelled it too narrowly, or placed emphasis on the wrong aspects of it. We then might wonder: what happens if our development of AI is informed by wrong notions of what we actually mean by ‘intelligence’ – worst still, by wrong ideas about human intelligence! A test bed for thinking and acting on this could be the creative industries, because it is in the creative industries that we so often combine ‘creativity’ and ‘intelligence’ (however we might define those concepts).

One of the most influential (and relatively early) books on artificial intelligence is What Computers Can’t Do: A Critique of Artificial Reason, published by philosopher Hubert Dreyfus in 1972. Dreyfuswas critical of AI. To summarize faintly, his arguments largely challenged the ability of AI to emulate the mental functions of human beings. Naturally, the AI community – including those working at MIT, where he had once been based – did not all take well to his critique of the inabilities of AI, and What Computers Can’t Do became something of a cause célèbre around which future work was built, both for and against AI’s capacities. Interestingly, while Dreyfus dealt with such things as human behavior, game-playing and language and pattern recognition, his attention to creative thinking and the imagination was brief.

There is an identified difference between creativity and intelligence (as currently defined) – and any correlation between these is by no means direct or balanced, as Kyung Hee Kim’s work shows, and as is incredibly important in the creative industries. This difference is notable in our consideration of how to develop Artificial Creative Intelligence (ACI). Significantly, the creative industries are also predominantly experience industries (involving observation, participation, skills, encounters, events) – so the role of the creative industries in generating as well as using Artificial Creative Intelligence (ACI) is heightened both in relation to material goods/products and in relation to human services/experiences. However, in order to avoid a misnomer do we now need to urgently revisit our ideas about what human ‘intelligence’ actually entails?

Misnomers often come about because something has been named before it was better understood. Sometimes misnomers arise to our advantage, because we revisit and redefine what we know; and, in our redefinitions, get closer to the truth. It is in this we might use Kim’s findings in conjunction with the active development of Artificial Creative Intelligence in the creative industries to revisit what we mean by human ‘intelligence’; and, in doing so perhaps get closer to the truth of what it actually entails.

While in 1972, Hubert Dreyfus’s critique challenged AI’s ability to emulate human mental functions, perhaps his critique needed to focus more on our creative functions – because it turns out (we might speculate) that it is these that most define human intelligence. If this is so, the creative industries can be the primary site not only for developing AI (which is already clear, given AI’s considerable and increasing impact on digital media, on publishing, on photography, on leisure software, on music and so on) but also of our investigation of the true nature of human intelligence.

Graeme Harper
[email protected].