Hank Green found the AI problem that YouTube labels can’t catch

Hank Green, co-founder of the educational YouTube channel Crash Course, has identified a potential flaw in YouTube's policies regarding AI-generated content. According to current guidelines, creators must disclose when they use AI to significantly alter or generate photorealistic material. However, this rule seems to have some inconsistencies, as it applies to "AI-generated music" but not to "riding a unicorn through a fantastical world," even though the latter could easily be photorealistic.

The YouTube policy further explains that "realistic AI content and meaningful changes require disclosure, while non-realistic or minor edits don’t." This statement creates confusion as it allows for a wide range of AI uses without mandatory disclosure. Examples given in the policy include everything from "idea generation" to "production assistance," which could range from using AI to create a video's outline to generating or enhancing a video's thumbnail, title, or infographic.

Interestingly, creators are allowed to use AI-generated or altered animation of a missile within a fully animated video without any disclosure. Additionally, creators can freely clone their own voices for voiceovers in their videos. Despite these broad and seemingly permissive guidelines, Hank Green has pointed out that there might be gaps in YouTube's policy, which could potentially lead to unregulated use of AI in video content.