"Don't Treat It Like Public Domain Material" - Disney Issues Cease and Desist to ByteDance: Where Are the Boundaries of AI Learning?

"Don't Treat It Like Public Domain Material" - Disney Issues Cease and Desist to ByteDance: Where Are the Boundaries of AI Learning?

1) What Happened: A Declaration of "Virtual Smash & Grab"

Disney, one of Hollywood's largest IP holders, has taken a hard stance against ByteDance, the parent company of TikTok. The focus is on ByteDance's AI video generation tool, "Seedance 2.0." Disney has sent a cease-and-desist letter, claiming that the tool is "unauthorizedly learning, generating, and distributing" Disney's popular characters and elements.


The symbolic expression used in the letter is noteworthy: "ByteDance's 'virtual smash-and-grab' against Disney IP is deliberate, rampant, and utterly unacceptable." By framing the issue as "theft" rather than a mere "concern," the premise of negotiations shifts.


2) The Core of Disney's Argument: "Don't Treat It Like Public Domain Material"

According to reports, Disney is concerned that Seedance includes Disney characters as if they were "public domain clip art," available for use without permission.


Furthermore, specific examples such as Spider-Man, Darth Vader, Grogu (Baby Yoda), and Peter Griffin from 'Family Guy' are cited as assets under Disney's umbrella. This is not a debate about "resemblance" but rather a clear presentation of "recognizable IP," aiming to converge the discussion on the "legality of learning" and "infringement of generated content."


A painful point is that "the infringement has already been distributed on social media and spread as a demo." Generated AI content doesn't end "inside the box," as users post it for viral purposes, exponentially increasing the scale and speed of infringement. Disney portrays this as "widespread infringement."


3) "Just the Tip of the Iceberg in Days"—The Reality Presented by Seedance 2.0

The letter includes a statement to the effect that "even though it was released just a few days ago, this is just the tip of the iceberg."
This phrase highlights the dual scale problem inherent in AI generation.

  • Scale of Learning Data: Feeding tens of millions or hundreds of millions of data points makes it difficult to "fully track" what is mixed in.

  • Scale of Distribution: Generated infringing content spreads exponentially through posting, reposting, clipping, and re-editing.


In other words, even if ByteDance makes "good-faith adjustments," stopping "already disseminated generated content" and the "prompt culture that induces similar generation" becomes a separate issue.

4) Hollywood's Encirclement: The Significance of MPA and SAG-AFTRA Acting Simultaneously

It's also important that this issue isn't just "Disney's fight." The industry group MPA (Motion Picture Association) criticizes Seedance for launching the service "without meaningful safeguards against infringement" and insists that the infringing activities should be stopped immediately.


Additionally, the actors' union SAG-AFTRA has issued a strong official statement, condemning the "unauthorized use of voices and likenesses" as "unacceptable." This ignites issues not only of copyright but also of **publicity rights, likeness rights, and consent**.


5) Reactions on Social Media: Divided Opinions on "Creativity" and "Exploitation"

Reactions on social media are divided into four main categories.

 


A. Creator Advocates: "This is 'Exploitation,' Not 'Learning'"

The narrative is that advancing in a direction that absorbs and replaces human creations without permission is a destruction of culture. Statements from organizations advocating responsible AI use strong language, asserting that "all legal means should be used to stop 'wholesale theft.'"


B. Tech Optimists: "Generated AI Can't Be Stopped. So, Make Rules Quickly"

The stance is that "since it can't be stopped, comprehensive licenses and distribution models should be created to 'legalize' it." This perspective seeks practical solutions based on technological progress. The Guardian's report also discusses the need for negotiations between AI companies and the creative side.


C. "Freedom of Expression" Advocates: "Will Fan Art and Parodies Be Restricted Too?"

There is backlash against Disney's strong stance, with concerns that "secondary creation culture will shrink." Here, the legality of "learning" and the distribution of "generated content" can easily become entangled, leading to heated debates.


D. Legal and Regulatory Watchers: "Next, It Will Tie Into TikTok Regulations and Sino-American Tensions"

Given that the opponent is ByteDance, posts discussing the issue in the context of geopolitics and regulation are increasing. The copyright discussion easily connects to debates on data crossing borders and platform governance.


Notably, Engadget's post has been widely shared as a source of news, with discussions expanding through quoted reposts.
Meanwhile, it can be confirmed that the relevant Reddit thread had little reaction at least immediately after its creation (in a "waiting for the first comment" state).


6) Where Are the Issues Headed? The Three Stages of "Learning," "Generation," and "Distribution"

The complexity of this issue lies in the fact that there is not just one point of contention.

  1. Learning Stage: Is it legal to include copyrighted works in learning data (discussions of fair use, etc.)?

  2. Generation Stage: Does the generated result constitute "reproduction or adaptation of characters"?

  3. Distribution Stage: When it spreads on social media, who is responsible and to what extent (users/platform/providers)?


Disney's strategy this time is to present these three stages collectively as "business infringement." Moreover, the claim that "this is the state in just a few days" reinforces the impression of "market entry with unprepared safeguards."

7) The Ironic Point: Disney Does Not Rule Out Partnering with AI

Importantly, Disney is not outright rejecting AI itself. According to Axios, Disney may be willing to partner with AI companies under certain conditions.


In other words, the issue is not "AI or anti-AI," but rather converging on **"unauthorized or authorized (and the associated compensation and control)."**


Source URL

Related Articles