Creative AI tools increasingly allow for the partial or complete automation of creative tasks. Be it through the augmentation of existing creative software or through embedded real-time generation; these algorithms have a growing influence on creative practices. Creative AI will not take over the world, but it certainly increasingly impacts creators, students, educators, and the industry at large.
Now that generative algorithms have human-competitive skills for many creative tasks and are being deployed for professional and amateur alike, it is critical to evaluate and discuss the implications of such developments. We introduce challenges and opportunities arising through a series of examples of generative systems developed at the Metacreation Lab and experiments conducted with these systems. In response to the polarization against the corporate big (and often stolen) data prompting tools sold to us as universal and omnipotent, we will expose the limitations of these approaches.
While doing so, we will introduce a range of alternative and ethical approaches readily available for computer-assisted sound design, music composition, and visual generation. We will advocate for small data and model crafting that can re-empower artists while capturing the cultural and aesthetic specificities lost with Big Data. We will present results of evaluations conducted with artists, students, and the software industry focusing on user experience, ethics, sense of authorship, phenomenology, and technological acceptance. Throughout, we will insist on the need for awareness and literacy on AI diversity, to counter the dominant corporate discourse, and help avoid further polarization of the cultural sector.
Philippe Pasquier is a Professor at the Simon Fraser University School of Interactive Arts and Technology (SIAT), where he directs the Metacreation Lab for Creative AI. Working at the intersection of generative AI, computational creativity, human–AI co-creation, and participatory creative AI systems, the Lab develops methodologies for human-centered AI design that emphasize co-creation, direct manipulation, and user agency rather than automation and replacement. Their research-creation practice spans music, moving image, and interactive media, with work presented internationally in venues such as Mutek, Ars Electronica, Centre Pompidou, and the Sydney Biennale. Philippe has co-authored 200+ peer-reviewed publications (9 best-paper awards), teaches a widely followed MOOC on generative art and computational creativity, and collaborates with the creative community and software industry to bring co-creative tools into real-world workflows.