After the trailer for Twins Himahima, an anime project with AI involved in 95% of its production process, was released, the anime received massive attention along with criticism, claiming that the anime has low quality and its only intention was to make money. Soon to be released in April this year, Twins Himahima demonstrates how AI can be used in animation while also revealing many potential underlying issues (Otaku, 2024).

For generative AI (GenAI), problems arise from the issue of copyright, and as the technology develops, a guideline that balances creativity, originality, and efficiency is needed. GenAI is trained on a vast dataset, and generates outputs based on given prompts; it can create human-like workpieces by identifying and labeling elements from human artists’ works. GenAI relies on a machine learning model called deep learning, which allows it to use algorithms to generate output similar to that of the human brain (Stryker et al., 2024).

In the animation industry, GenAI is used not only in image or text generation but also in 3D models and videos creation. Through deep learning, GenAI can comprehend real-world physics, such as gravity and collision dynamics, from inputting sources to better interpreting visual data and generating output that accurately depicts interactions between objects based on annotated footage (Ai-Admin, 2024). DeepMotion is one such GenAI that studies object physics, transforming real-world motion into 3D animation and generating lifelike jumps or spins (Ai, 2024).

GenAI is trained on data collected from diverse sources, such as internet scraping and crawling, books, scientific journals, free image libraries, and news articles (Samir, 2023). Data labeling is a crucial phase in training GenAI, involving identifying and adding labels that help the model recognize objects and understand expected outputs (Casey, 2023). For example, GenAI creates in-betweens in animation—the intermediate frames between keyframes. During this process, labeling is required to mark the start and end points of a motion and categorize facial expressions. 

Many animation companies, such as Disney, use GenAI and AI-assisted technology. As one of the world’s largest conglomerates, Disney has faced the pressure of producing visuals efficiently while maintaining its high quality, which urged them to proactively incorporate GenAI in their animation, such as creating in-betweens, advancing rendering, promoting rigging, empowering style transfer, and so on (DigitalDefynd, 2025). For instance, to empower style transfer, AI analyzes images and makes the photo more realistic or simulates different lighting conditions (Fox Render Farm, 2024). Disney established a new Technology Enablement Office (Ferguson, 2024) and designed a new GenAI tool for robotics called Newton, which brings characters to life with smoother motion via its new physics engine (Burias, 2025).

In Japan, AI is also prevalently used in the animation industry. Toei Animation, one of Japan’s oldest animation studios, began using AI to speed up its production process, such as converting the photos of cities into images with anime-like rendering (De Wit, 2023). For example, in the short video, Urvan, Toei used AI to convert photos of Japan’s city Sasebo into pink-hued, cyberpunk, animated-like backgrounds in the video, demonstrating the surging trend of incorporating AI in animation production. 

Revolutionize the Animation Industry

Using AI-assisted technology in animation has two main benefits: improving efficiency and fostering innovation.

For animation companies, GenAI and AI-assisted technology improve efficiency due to their ability to create output nonstop and much faster than humans. AI DeepMotion users have reported that GenAI simplified the animation process, which would have taken weeks to do without (Ai, 2024). Moreover, AI-assisted technology can improve the overall quality of work, making the animation smoother by adding in-betweens with more details from the learning process. The director of To Dear Me, the 2024 Beijing International Film Festival’s best anime film, stated that using AI largely shortened the creation time and enhanced finer details, such as adding details to the dandelion fluff, which would have taken a huge amount of time if it was hand drawn (Ding et al., 2024).

Reduced production time and improved quality of work allow production studios to focus more on creating emotional depth and conveying meaning, enhancing the overall innovative aspect of work. For example, Liu Yu, General Manager of Tencent Online Video’s Smart Creation and Content Platform Department, alleged that AI alleviates the artists’ burden of doing repetitive work and empowers artistic expression, allowing them to focus more on the creativity aspect (“Tencent Online Video Liu Yu: AIGC delves into film and television production to assist creative teams in producing high-quality content”, 2025).

The usage of AI is also cost-saving. DreamWorks founder Jeffrey Katzenberg predicts that with the help of AI-assisted technology, the production cost will be cut by 90% (Weprin, 2023). The budget saved from production and labor costs can be used in other fields, which can also increase the overall quality of animation (Connolly, 2025). For example, in To Dear Me, the dandelion appeared to be a white ball in the hand script, and AI added the detail of the fur texture on the dandelion (Ding et al., 2024).

For animation artists, GenAI makes it easier to create and can assist them in finding inspiration, brainstorming, and visualizing their thoughts. Trained by large datasets, GenAI can help artists visualize their ideas, enabling them to pick out the best way to deliver a certain idea and thus increasing their efficiency.

In conclusion, AI implementation is effective and time-saving, freeing artists from repetitive work to focus more on creativity, like storyline and character development. However, there are also potential drawbacks to using AI in animation production.

Beyond Efficiency: AI’s Double-Edged Impact and Violation of Privacy, Copyright, and Labor Right

To start with, GenAI training datasets in animation violate artists' copyright and labor rights.

Copyright protects all original works in a tangible form, including novels, art, poetry, musical compositions, graphic designs, and so on, but a registered copyright is needed if legal actions are taken (Kenton, 2024). In the AI era, many artworks are fed into the GenAI dataset without the artists’ consent. In 2023, Stability AI and other GenAI platforms were sued by artists because they used many artists’ original works to train their models and develop their styles—similar to the original artists—without a license. Additionally, AI has been trained with data from millions of unlicensed works, infringing artists’ intellectual property and copyright (Adi, 2024).

Besides violating copyrighted works, GenAI also blurs the boundaries of originality, intensifying preexisting issues regarding originality in the field. The U.S. Copyright Office ruled that copyright will not be given to works solely generated by AI but to those with originality, even if it is AI-assisted, as long as it reflects the author’s thoughts. Courts will rule such cases based on investigating the motivation behind a particular stroke the artist chose to use (Bitlaw, 2025). However, the extent to which AI should contribute to such co-produced work remains contentious. Thus, using AI can lead to unintentional excessive borrowing, which means GenAI can possibly generate a product that is similar to that of the works in its dataset without the user knowing. 

Furthermore, an artist’s style—often seen as essential to the originality and uniqueness of their art—is not protected by copyright law, meaning that GenAI can easily copy the style of a human artist through style transfer and create an output based on the content image and the style reference image (GeeksforGeeks, 2023). Thus, AI-assisted or generated works can copy other artists’ styles, while the artists of the original reference image are unable to sue those who copy their styles because the style is not legally protected. Such cases result in a deteriorating creative environment and devaluing originality (Blaszczyk et al., 2024).

Moreover, a reduction in labor costs causes unemployment and income inequality. Human artists who are more engaged in repetitive work will have less negotiation power because machines can easily replace them (Weprin, 2023).

A Delicate Balance and Existing Actions

Due to the existing benefits and drawbacks, a balance needs to be drawn. Using GenAI seems to be an inevitable trend in the animation industry due to its efficiency, yet it violates the rights of artists and workers involved in the field and creates an environment where copyright becomes even more vulnerable. Thus, while employing AI-assisted technology becomes unavoidable, it should be restricted.

Efficiency in animation production is important as the market demand for high-quality animation is rising. Thus, for animation companies like Disney, working with GenAI becomes one of their best resorts, and increased efficiency allows them to create more pieces of work in a limited time, allowing them to gain more profit while maintaining the quality of their work, and thus remaining competitive in the fast-changing industry (DigitalDefynd, 2025). Yet, despite increased efficiency, protecting copyright is essential since the industry can only develop healthily when individual rights are valued. Humans need to be at the heart of any technology development in order to protect their rights while the industry grows, as Volker Türk, UN High Commissioner for Human Rights, once said (Volker Türk, 2023). 

Some recent regulations address AI-assisted works. The U.S. Copyright Office (USCO) has been tackling copyright issues and preparing to update the Compendium of USCO Practices—the administrative manual for copyright registration. While dealing with copyright lawsuits, the office is also publishing a paper on the key debates regarding AI-assisted work’s legitimacy as original and eligible for copyright. Despite all the efforts, the criteria for AI-assisted pieces to be copyrighted are still vague since the paper does not address the ambiguity around GenAI. Creators who use GenAI in their creations might still be unaware of potential plagiarism, even though the artworks may be based on their original thoughts. The USCO gave an example of an AI-assisted art piece that will be given copyright registration. It is an image showing a woman with flowers coming out from her head. The final work was generated by GenAI with a hand-drawn input. However, while the input the user draws contains different types of flowers, the outcome generated with the GenAI only includes roses. Dissidents claim plagiarism due to overborrowing the work of human artists, who often use roses to depict a romantic sense of feeling (Bitlaw, 2025).

EU’s recent AI Act also tries to address the issues of AI and datasets usage. The Act considers different scenarios. It states that artist cannot prevent their work from being used by research organizations and cultural heritage institutions, but the security and integrity of their work shall be protected. In this way, even though the artist cannot prevent their works from being included in those non-profit datasets, they should know that their work will not be misinterpreted or used for malicious purposes. In cases of non-research commercial usage, the artists can reject the requests for their work to be fed into such datasets. This Act is effective because it protects copyrighted works from being included in the dataset for commercial purposes. However, even though copyright laws shall protect all original works, strong protection is only given to registered work, yet some artists may not be able to register all their work due to the long registration process and registration fees. Thus, when their works are integrated into the dataset without their consent, it is still hard for them to claim their due rights since a registering their infringed work is needed to sue a company (Blaszczyk et al., 2024).

Conclusion

The animation industry is at a turning point regarding the dilemma between efficiency and human rights. Not only animation, but all industries are adapting to the ever-evolving advancements of AI. Thus, a more well-developed and respectful future is expected to ensure technology is developed and used ethically and sustainably.

References

Ai-Admin. (2024, January 10). “Artificial Intelligence and the Dependence on Physics - Unraveling the Deep Connection.” AI For Social Good. aiforsocialgood.ca/blog/artificial-intelligence-and-the-dependence-on-physics-unraveling-the-deep-connection.

Alex, Ai. (2024, November 27). “DeepMotion AI Review: Transform 2D Videos Into 3D Animations With Ease - AI 3D Model Generation.” AI 3D Model Generation. ai3dmodelgeneration.com/deepmotion-ai-review.

Blaszczyk, Matt, et al. (2024, November 20). “Artificial Intelligence Impacts on Copyright Law.” Rand.org. www.rand.org/pubs/perspectives/PEA3243-1.html#:~:text=Copyright protects only original human-authored works, including those,interests of rights holders and the creative industry.

Burias, A. (2025, March 21). “Disney to Create Star Wars-Inspired Robot 'Newton' Using AI with Nvidia, Google”. Designrush. https://www.designrush.com/news/disney-starwars-inspired-robot-newton-using-ai-nvidia-google-deepmind.

Cole Stryker et al. (2024, March 22). “What is generative AI?” IBM.com. www.ibm.com/think/topics/generative-ai.

Connolly, Michelle. (2025, February 23). “Animation and AI: Revolutionize Creativity in Filmmaking.” Educational Voice. educationalvoice.co.uk/animation-and-ai/#enhancing-the-creative-process-with-ai.

“Copyright Protection for AI-Generated Works”. (2025). Bitlaw. https://www.bitlaw.com/ai/AI-copyright.html.

De Wit, Alex Dudok. (2023, February 14). “Japan’s Toei Is Using AI Tools to Speed up Animation Production.” Cartoon Brew. www.cartoonbrew.com/anime/japans-toei-is-using-ai-tools-to-speed-up-animation-production-206801.html.

DigitalDefynd, T. (2025). 5 Ways Disney is using AI [Case Study][2025] - DigitalDefynd. DigitalDefynd. https://digitaldefynd.com/IQ/ways-disney-use-ai/

Ding Yazhi et al. (2024, October 26) “Will AI revolutionize the industry of animation?”. Tencent News. https://news.qq.com/rain/a/20241026A00X7U00

Ferguson, Mackenzie. (2024, November 2). “Disney Unveils Office of Technology Enablement to Spearhead AI and MR Innovations.” opentools.ai. opentools.ai/news/disney-unveils-office-of-technology-enablement-to-spearhead-ai-and-mr-innovations.

GeeksforGeeks. (2023, December 6). Neural Style Transfer with TensorFlow. GeeksforGeeks. https://www.geeksforgeeks.org/neural-style-transfer-with-tensorflow/

Kaplan, Soren, PhD. (2023, December 28). “9 Ways AI Can Unlock Your Creativity in Life and Business.” Psychology today. www.psychologytoday.com/us/blog/the-power-of-experience/202312/increase-your-creativity-with-artificial-intelligence.

Kenton, Will. (2024, August 12) “Copyright: Definition, Types, and How It Works.” Investopedia. www.investopedia.com/terms/c/copyright.asp.

Matty Casey. (2023, September 29). “Data labeling: a practical guide (2024)”. Snorkel.ai. https://snorkel.ai/data-labeling/

Otaku, Sweety. (2024, December 15). “Twins Hinahima, an Anime Made With AI, Is Announced” Anime Sweet. animesweet.com/anime/twins-hinahima-an-anime-made-with-ai-is-announced.

Robertson, Adi. (2024, August 13)“Artists’ Lawsuit Against Stability AI and Midjourney Gets More Punch.” The Verge. www.theverge.com/2024/8/13/24219520/stability-midjourney-artist-lawsuit-copyright-trademark-claims-approved.

Sampat Samir. (2023, September 20). “Where Do Generative AI Models Source Their Data and Information?” Smith.ai. smith.ai/blog/where-do-generative-ai-models-source-their-data-information.

Shah, Suvrit. (2025, March 21). “AI For Animation: From Storyboarding to Final Render – Boost Efficiency &Amp; Creativity.” Yeti AI. yetiai.com/ai-for-animation-from-storyboarding-to-final-render.

“Tencent Online Video Liu Yu: AIGC delves into film and television production to assist creative teams in producing high-quality content” (2025, February 22). AIGC. https://www.aigc.cn/88159.html

Volker Türk, (2023, July 12) “Artificial intelligence must be grounded in human rights, says High Commissioner”. Ohchr.org. https://www.ohchr.org/en/statements-and-speeches/2023/07/artificial-intelligence-must-be-grounded-human-rights-says-high

“What Data Is Used to Train an AI, Where Does It Come From, and Who Owns It?” (“no date”) Potter Clarkson. www.potterclarkson.com/insights/what-data-is-used-to-train-an-ai-where-does-it-come-from-and-who-owns-it.

Weprin, Alex. (2023, November 9). “Jeffrey Katzenberg: AI Will Drastically Cut Number of Workers It Takes to Make Animated Movies.” The Hollywood Reporter.  www.hollywoodreporter.com/business/digital/jeffrey-katzenberg-ai-entertainment-animation-prediction-1235643311.

“What is AI Rendering? How AI is Revolutionizing 3D Animation Production”. (2024, December 11). Fox Render Farm. https://www.foxrenderfarm.com/news/what-is-ai-rendering/.