21 August 2014 in Education and Information Technologies
24 August 2018 in Connection Science
In human–computer co-creativity, humans and creative computational algorithms create together. Too often, only the creative algorithms and their outcomes are evaluated when studying these co-creative processes, leaving the human participants to little attention. This paper presents a case study emphasising the human experiences when evaluating the use of a co-creative poetry writing system called the Poetry Machine. The co-creative process was evaluated using seven metrics: Fun, Enjoyment, Expressiveness, Outcome satisfaction, Collaboration, Ease of writing, and Ownership. The metrics were studied in a comparative setting using three co-creation processes: a human–computer, a human–human, and a human–human–computer co-creation process. Twelve pupils of age 10–11 attended the studies in six pairs trying out all the alternative writing processes. The study methods included observation in paired-user testing, questionnaires, and interview. The observations were complemented with analyses of the video recordings of the evaluation sessions. According to statistical analyses, Collaboration was the strongest in human–human–computer co-creation, and weakest in human–computer co-creation. Ownership was just the opposite: weakest in human–human–computer co-creation, and strongest in human–computer co-creation. Other metrics did not produce statistically significant results. In addition to the results, this paper presents the lessons learned in the evaluations with children using the selected methods.
2 January 2019 in Digital Creativity
Collaboration with artificial intelligence (AI) is a growing trend even in the field of creativity. This paper examines which quantitative metrics can be used to comparatively analyse human-computer co-creativity with children. To study this question, 24 schoolchildren of age 10–11 wrote a poem with three co-creative poetry writing processes: a human-computer, a human-human, and a human-human-computer process. The computational participant in the processes was an AI-based application called the Poetry Machine. The children were asked to evaluate their user experience with a 5-point Likert-type questionnaire after each writing process and a comparative questionnaire after finishing all processes. The metrics used in the evaluation were immediate fun, long-term enjoyment, creativity, self-expression, outcome satisfaction, ease of starting and finishing writing, quality of ideas and support from others, and ownership. Significant differences were found in fun, long-term enjoyment, quality of ideas, support, and ownership. The high number of statistically relevant results was enabled by exposing all participants to all writing processes, and the comparative questionnaire. The human-human-computer process was evaluated the best in long-term enjoyment and the human-computer process the weakest in support and idea quality. Creativity and ease of finishing writing turned out to be outlining metrics for the co-creative processes.