YouTube's AI Moderation Unfairly Targets VTubers Leaflit, SmugAlana, and Kitsu

Rev says desu

Introduction: YouTube's AI Moderation Issues [0:00] [Image 0]

The video opens by addressing the ongoing issues with YouTube's platform, particularly how its AI moderation negatively impacts content creators, especially VTubers. Many creators are experiencing demonetization due to their content being flagged as "inauthentic" by YouTube's AI systems. The speaker notes a lack of media coverage on these widespread problems. [0:37] [Image 1]

VTuber Leaflit Receives Strike for "Harassment and Cyberbullying" [1:32] [Image 2]

A specific case is presented involving VTuber Leaflit, who received a strike on her YouTube channel due to a video being removed under the "Harassment and Cyberbullying Policy." [1:39] [Image 3] [1:42] [Image 4]

• Context of the Strike [2:14] [Image 5]
  • Leaflit's tweet to Team YouTube explains the strike was for "content that targets someone with prolonged insults or slurs based on their intrinsic attributes."
  • The specific incident involved a clip where parent Simone H Collins affectionately called her children "little bastards" as a nickname while playing.
• Critique of YouTube's Policy Application [3:06]
  • Leaflit argued that "policing how parents express love to their children" feels deeply dystopian, comparing it to common affectionate terms like "little monsters."
  • The speaker emphasizes that the nickname was clearly affectionate and not meant to demean, suggesting the AI moderation or an "overzealous moderator" made a clear error.
  • The speaker highlights the difficulty in getting a human review or correction from YouTube, often requiring public outcry on platforms like Twitter.

VTuber SmugAlana's Reach and Impressions Drop to Zero [5:18] [Image 6]

Another VTuber, SmugAlana, experienced a significant and sudden decline in her channel's performance.

• Impact on SmugAlana's Channel [5:30]
  • On April 28th, SmugAlana's reach and impressions plummeted to zero overnight, affecting all her videos, with no strikes or warnings issued.
  • She questioned if she was caught in an AI "inauthentic content" sweep due to being a VTuber, showcasing analytics that illustrate a sharp, non-organic drop in views.
• Team YouTube's Response [6:53]
  • Team YouTube responded to SmugAlana's public tweet, sending her a direct message to "get more details & figure this out together." The speaker expresses skepticism about the effectiveness of a DM for such a deep-seated issue.

VTuber Kitsu Faces Similar Issues and a Dismissive Response [7:17]

The video presents a third case involving VTuber Kitsu, who also reported similar negative impacts.

• Kitsu's Experience with Suppressed Content [7:18]
  • Kitsu tweeted about experiencing "same BS" with dropped impressions and reach, being in a "3-day jail" where her videos were not getting suggested. She stated this had been happening for two months.
• YouTube's Unprofessional Reply [7:40]
  • Kitsu shared an email response from YouTube Partner Support which stated her channel was "fully investigated and it's working as intended."
  • The email controversially added, "If you are not satisfied with having 2,814,469 impressions despite having only 94.1K subscribers in your channel, report to our technical team," which the speaker labels as "passive-aggressive and unprofessional."
  • The speaker posits that this response, due to its tone and specific numbers, likely came from a human reviewer, indicating a lack of understanding or empathy within YouTube's support.

Broader Implications of YouTube's Algorithm and Moderation [9:06] [Image 7]

The discussion expands to the larger systemic issues creators face with YouTube's moderation and algorithms.

• Automated Scoring Systems and Channel Siloing [9:09]
  • The speaker references a "supposed rumor" about an automated scoring system that, once a channel crosses a certain threshold for "hidden spam or dangerous content," silos it to a smaller and smaller group of viewers.
  • This often precedes demonetization or a channel ban, highlighting an opaque process that can severely harm channels without clear warnings.
• The Broader Pattern of Platform Challenges [10:07]
  • The speaker likens navigating YouTube to "flying a plane" where parts randomly stop working, such as notifications, video discoverability, or forced unsubscribes.
  • This indicates an unpredictable environment for creators, especially those covering "anything controversial," suggesting that YouTube prioritizes "reliable sources of income" like Hollywood and celebrities.
• Call for Transparency and Accountability [11:47]
  • SmugAlana advocates for a data analysis of various YouTubers (VTubers and non-VTubers) to identify correlations and patterns in these issues.
  • The speaker concludes that YouTube is "dropping the ball" with its increasing reliance on AI moderation, leading to more problems and less transparency, urging creators to continue speaking out and interacting on social media to draw attention to these issues.