ADVERTISEMENT

OpenAI Faces Copyright Violation Risk For Sora 2 Training On Studio Ghibli, Japanese Artwork

AI-generated images in the style of Studio Ghibli, produced via OpenAI’s ChatGPT, have gained popularity online for replicating the studio’s distinctive aesthetic.

<div class="paragraphs"><p>A group supported by Studio Ghibli has called on Sam Altman's OpenAI to halt the incorporation of Japanese artwork in the training of its Sora 2 AI model. (Source: Sam Altman/X)</p></div>
A group supported by Studio Ghibli has called on Sam Altman's OpenAI to halt the incorporation of Japanese artwork in the training of its Sora 2 AI model. (Source: Sam Altman/X)
Show Quick Read
Summary is AI Generated. Newsroom Reviewed

A group supported by Studio Ghibli has called on OpenAI, led by Sam Altman, to halt the incorporation of Japanese artwork in the training of its Sora 2 AI model.

The Content Overseas Distribution Association (CODA) submitted a letter to OpenAI on Oct. 28, expressing worries regarding the firm’s video-generation tool, Sora 2 (available at sora.chatgpt.com), which debuted on Sept. 30.

The prominent Japanese trade body, which includes major publishers and Studio Ghibli among its members, has officially requested that OpenAI discontinue the use of Japanese copyrighted materials in training its AI systems. This plea was issued last week.

In the past, AI-generated images in the style of Studio Ghibli, produced via OpenAI’s ChatGPT, have gained popularity online for replicating the renowned studio’s distinctive aesthetic. Yet, Hayao Miyazaki, co-founder of Studio Ghibli and a vocal opponent of AI in art, has stated that these pictures contradict the principles of true artistry.

OpenAI Faces Risk Of Copyright Violation For Sora 2 Training

CODA noted that it has verified a significant amount of Sora 2’s output bears a strong similarity to Japanese content or visuals, suggesting that Japanese creations were likely utilised as training data for the AI.

CODA contended that duplicating protected works in the AI training phase might constitute infringement, particularly if the resulting outputs mimic or copy pre-existing artistic pieces.

The group also criticised OpenAI’s opt-out mechanism, which permits rights holders to seek exclusion of their content from datasets post-training. According to CODA, this method runs counter to Japan’s copyright laws, which typically mandate advance approval for using safeguarded materials.

The letter from CODA specified two primary requests:

  • In the operation of Sora 2, CODA requests that its members’ content is not used for machine learning without their permission.

  • OpenAI responds sincerely to claims and inquiries from CODA member companies regarding copyright infringement related to Sora 2’s outputs.

Opinion
Meta Rejects Allegations Of Using Porn For AI Training: 'Videos Were For Personal Use'

Recent Backlash Over Sora

CODA’s objections come on the back of OpenAI encountering backlash over its management of delicate subjects in Sora. The company recently suspended the creation of images featuring Martin Luther King Jr., in light of protests from the civil rights icon’s estate.

On Oct. 17, OpenAI shared a statement on X, explaining that it collaborated with The Estate of Martin Luther King, Jr. Inc. to handle the leader’s pictures with dignity. 

“At King, Inc.’s request, OpenAI has paused generations depicting Dr King as it strengthens guardrails for historical figures,” the AI company said.

Opinion
OpenAI Clarifies ChatGPT Policy: No Change In Legal Or Medical Advice Rules
OUR NEWSLETTERS
By signing up you agree to the Terms & Conditions of NDTV Profit