Stakeholder Groups in Computational Creativity Research and Practice - Computational Creativity Research: Towards Creative Machines - page 14

1.4.3 Software Accounting for Its Actions
We argue in [ 32 , 40 ] that people take into account how a person or software operates
when they assess the value of the output it produces. To address this issue, we advocate
a development path to follow when building creative software: (i) the software is
given the ability to provide additional, meta-level, information about its process and
output, e.g., giving a painting or poem a title; (ii) the software is given the ability
to write commentaries about its process and its products; (iii) the software is given
the ability to write stories—which may involve fictions—about its processes and
products; (iv) the software is given the ability to engage in dialogues with people
about what it has produced; how and why. Indeed, giving software the ability to
discuss its creative works would mirror Turing's original proposal for an intelligence
test [ 34 ] to a greater extent than tests focusing only on consumer perception of
artefacts. As a preliminary example, in [ 41 ], we demonstrated a poetry generation
system which is able to provide commentaries about its poetry, and how and why it
produced a particular poem.
As we discuss in [ 42 ], in a computational setting, there are advantages to software
being immersed in environments where serendipity might occur. However, account-
ing for lucky events that trigger creative acts may actually lessen the celebration
and hence the impact that the acts have. It is important to note that people tend not
to describe their processes and products in the explicit way we advocate for soft-
ware, preferring to maintain some level of mystery. Nevertheless, we believe that,
at this stage in the development of computationally creative systems, it is important
to address the humanity gap—without aspiring to eliminate it. Framing [ 40 ]serves
to highlight that intelligent processing was used to produce artefacts, which is an
important first step. Given that audience members will typically not be able to come
up with an interesting backstory without some scaffolding, positive acts of framing
are likely to have more fruitful impact than an overall air of mystery.
Another possible way to address the humanity gap is to manage people's expec-
tations about the level of humanity they will encounter through a computationally
produced artefact. In the sameway that when people buy an e-book they know they are
not going to get a physical object, we advocate telling audiences that they are reading
a c-poem , and hence—in the knowledge that it was produced computationally—they
will get a reduced human connection. We can go further in re-imagining traditional
artefacts, for instance in suggesting that a c-poem is actually a doublet of texts, one
which resembles a traditional poem and another which provides a commentary about
the motivations, actions and results of the software's processing. We believe this will
highlight the humanity gap, but that it will do so in such a way as to help people
to engage with and appreciate the creative process, and better enjoy the artefacts
produced by software.
Hypothesis 4 The humanity gap can be addressed by re-imagining the nature of cre-
ative artefacts, to manage expectations of humanity. In particular, it is advantageous
for software to account for its processes and products through additional material
such as a commentary.