Agency and legibility for artists through Experiential AI

Art and experience can make AI more understandable: The paper demonstrates that experiential AI—using tangible, creative experiences rather than technical explanations—helps make AI systems "more transparent, and therefore accountable" by bridging the gap between complex code and human comprehension.

Artists need more control over AI tools: The case study shows that when artists can explore and understand how AI systems work through hands-on experimentation, they gain "creative control over the AI process," suggesting that current AI tools often limit creators' agency and decision-making power.

Experiential AI increases both understanding and accountability: By making AI's "values and implications more transparent" through creative experiences, this approach addresses a key responsible AI concern—ensuring that people affected by AI systems can actually understand how they work and hold them accountable.

Interdisciplinary collaboration is essential for AI transparency: The paper emphasizes that "artists, scientists and other interdisciplinary actors" working together can develop better methods for communicating AI's "functionality, limitations, and consequences" than technical experts working alone.

Creative methods should complement traditional AI explainability: The research suggests that arts-based approaches can "add to the toolbox for human-centred XAI" (explainable AI), indicating that responsible AI development benefits from creative strategies alongside conventional technical explanations.