The Freelance League of Japan (hereafter FLJ) recently announced the results of a large-scale study on how Japanese creators view generative AI. The survey had a sample size of 24,991 people, including freelancers working in various creative fields. The most represented group were visual artists, with 71.3% of respondents being people whose work involves drawing, like illustrators (54.2%), manga artists (15%) and animators (2.2%).
As reported by Otaku Soken, 88.6% of total respondents (65.3% “strongly agree” + 23.3% “somewhat agree”) believe that generative AI poses a “serious threat” to their livelihood, and 93.3% expressed concern over losing future work or work they’re currently engaged in. Additionally, around 12% of Japanese creatives said that their earnings had actually decreased due to the impact of gen-AI. About 10% of respondents reported opting to secure alternative sources of income (which are not dependent on creative work).
Based on the survey, Japanese artists are strongly in favor (92.8%) of generative AI models being legally required to disclose a list of copyrighted works that were included in learning data. When it comes to giving consent for machine learning, the majority of creatives (61.6%) want prior permission to be a requirement (as opposed to the now commonplace “opt-out” system that assumes permission until its revoked). On the other hand, 26.6% of respondents would prefer for learning to be “prohibited in principle.”
When asked about potential revenue sharing systems for generative AI models, such as subscription- or license-based models, the majority answer (33.3%) was “I can’t agree with any option.”
Based on the survey results, the FLJ has made some recommendations to Japan’s government and relevant administrative bodies, like mandating transparency for training data, mandating labeling and establishing guidelines for whether output is illegal, creating a fair revenue-sharing scheme, developing labor policies, and establishing a governing body.
