This thesis examines stakeholder perceptions of artificial intelligence in film and television through a mixed-methods approach. The core analysis employs a MaxDiff utility measurement with 611 general audience members and 48 entertainment professionals to quantify acceptance patterns across 19 distinct AI use cases. Additionally, a question wording effect experiment reveals that describing technology as "AI" versus "software that scans automatically" results in a statistically significant difference of 12.77 percentage points, highlighting how terminology influences perception independent of functionality. When combined with interview insights, these findings reveal consistent psychological boundaries: non-creative assistance applications receive broad acceptance while creative autonomous applications face strong resistance, a tension between enhancing versus replacing human creativity. AI implementations that maintain human guidance dramatically outrank their autonomous counterparts across all demographics. Demographic analysis reveals age as the strongest indicator of AI acceptance, with younger respondents showing greater openness to creative applications while older participants demonstrate stronger acceptance of non-creative tools but heightened resistance to creative uses. Surprisingly, media consumption habits (self-identified viewer type) show minimal influence on acceptance patterns, suggesting that AI perceptions are rooted in deeper values rather than entertainment engagement levels. Industry professionals demonstrate more nuanced distinctions than general audiences, showing stronger protection of core creative domains while expressing greater appreciation for technical assistance. These findings offer practical guidance for implementation strategies that respect domain-specific boundaries while leveraging AI capabilities in ways that preserve human creative direction.