Stakeholder Groups in Computational Creativity Research and Practice - Computational Creativity Research: Towards Creative Machines - page 12

1.4.1 Essential Behaviours

We suggest not asking people if they believe software is behaving creatively, but
rather concentrating on whether they perceive the software to be acting uncreatively .
Using our standpoint above that the notion of creativity is essentially contested [ 3 ], we
expect that no matter how sophisticated our software gets, we will not see consensus
on such matters. However, we have found that people agree much more on notions
of uncreativity: if a program doesn't exhibit certain behaviours onto which certain
words can be projected, then it is easy to condemn it as being uncreative. Building on
the foundational arguments given in [ 32 ], we propose that audience members can too
easily label software as uncreative if they are unable to project any of the following
words onto the behaviours they perceive software to be exhibiting:

skill, appreciation, imagination, learning, intentionality, accountability, innovation,

subjectivity and reflection

We have found that assessing the level of projection of these words onto the
behaviours of software can help us to gauge people's opinions about (the lack of)
important higher-level aspects of software behaviour, such as autonomy, adaptability
and self-awareness. Note that we make no claim about the above behaviours being
sufficient for a perception of creativity: a necessary set of behaviour types for avoiding
the uncreativity label is not the same as a sufficient set of behaviour types for gaining
the creativity label. This mis-interpretation of our aims for highlighting the above
essential behaviours has propagated somewhat, for instance in [ 33 ].

Hypothesis 2 Creativity in people and software is essentially contested and sec-
ondary, and hence it might be advantageous to work on people's perception of
uncreativity in software, as this is easier to predict/manage. Software exhibiting
the essential behaviour types highlighted above is necessary for it to avoid being
labelled as uncreative. Eventually, when there are no good reasons to label software
as uncreative, people may choose to label it as creative.

1.4.2 The Humanity Gap

One could argue that, given the particularly human-centric nature of creativity, and
that a human connection is paramount in much of the arts, it is simply inappropriate
to use the term 'creative' to describe software. The status quo is that we currently
haphazardly apply human terminology related to creativity to software, which often
requires the projection of other human qualities onto software, such as it being juve-
nile, which is inherently error prone, given that computers are patently not people.
Another option is to ignore the non-human nature of software and concentrate on
what it produces, rather than on what it is, or what it does. To begin to address the
kind of silicon biases described above, researchers often compare the interpreta-
tion of computer-generated and human-produced artefacts in a rather extreme “blind