Anim[æ]ting Intelligence

Artificial Intelligence has become an ever-present phenomenon, its potential and risks causing excitement and fears. This thesis asks how AI’s effectiveness varies in creative workflows, specifically across the stages of anime production. It found trends suggesting that AI art may become its own genre and that a variety of ways to incorporate it into creative workflows exist. Exploring AI’s risks, the responsibilities of users, developers, and lawmakers were described, which are to be considered in relation to risks. Providing an overview of anime production and the state of its industry, the thesis shows unsuccessful attempts at using AI to aid it, including its negative reactions. In a self-experiment, a short anime film was produced using AI, examining the hypothesis that AI’s effectiveness will be inversely proportional to the level of human-centered emotional nuance of each task. Indeed, the variation of AI’s efficiency was great and largely followed the hypothesised trend.