How AI Agents Address Copyright Concerns

Set “Book Fragrance Shanghai” as a top star mark

Let the fragrance of books accompany you

How AI Agents Address Copyright Concerns

“The emergence of AI agents, the practice of ‘AI training supplementary agreements’, and the concerns of contracted writers on online platforms remind us to rethink and redefine the essence of creation. We need to leverage the value of artificial intelligence while affirming and protecting the primary status of human creators,” said Xing Zheng, Director of the Management Committee of Shanghai Jinshen Law Firm and Doctor of Law at Shanghai Academy of Social Sciences, in a recent exclusive interview.

Recently, the “Lingka Lingka” creative writing AI agent launched by the School of Communication at East China Normal University was released at Shanghai Book City. In recent years, AI-assisted writing tools represented by platforms such as the “Yuewen Miaobi” and “Writer Assistant Miaobi Version” from Yuewen Group and “AI Assistant” from Qimao Chinese Network have shown their capabilities. With the upgrade and optimization of technology, the rise of AI writing software on major platforms may indicate that human-computer collaboration will become a new trend in creative writing. Is a new era of writing about to begin?

How AI Agents Address Copyright Concerns

Intelligent tools, with their powerful algorithms and data processing capabilities, bring unprecedented convenience and efficiency to literary creation. However, the development of AI writing also faces new challenges. In July this year, the Tomato Novel platform attempted to include authors’ original content in corpus training, requiring the signing of an “AI training supplementary agreement”, which sparked resistance and complaints from some contracted writers. Such incidents not only reveal the issues of how to protect copyright and the rights of original authors in AI writing but also prompt us to rethink and examine the essence and ethical norms of literary creation in the age of artificial intelligence.

When AI Agents Meet Online Literature

AI writing software, through deep learning and natural language processing technologies, can simulate human writing styles and thought processes, providing authors with a wealth of creative materials and inspiration. A team led by Professor Wang Feng from the School of Communication at East China Normal University successfully completed a novel titled “The Apostle of Destiny” with over 1.1 million words in May this year using a combination of “domestic AI large language models + prompt engineering + human post-editing”. In October 2023, a team led by Professor Shen Yang from Tsinghua University’s School of Journalism and Communication created the novel “Memory of the Machine” using generative artificial intelligence, anonymously participating in the Jiangsu Province Youth Science Popularization Science Fiction Works Competition and winning the second prize.

What impact does AI writing agents have on online literature? Will AI writing software bring about a new writing model?

Currently, many online writers are using AI to assist in creation, such as conducting research, story brainstorming, scene imagining, and character portrayal. For online literature, which requires physical, intellectual, and endurance labor, the involvement of AI helps to accelerate writing speed. Suo Luo, a member of the Chinese Writers Association’s Online Literature Committee and an online literature critic, analyzed the unique expression methods and production models of online literature. He believes that the elements of “database writing” such as types, routines, and tropes in online literature have the inherent possibility of being dismantled and reorganized; at the same time, the meaning expressed in online literature is relatively superficial and straightforward, not emphasizing poetic “implications.” These factors provide convenience for AI writing. Essentially, AI writing still operates based on algorithms, mobilizing resources according to human-defined requirements to reorganize or generate new texts.

The works generated by AI remind us to rethink and redefine the essence of creation. Xing Zheng stated, “In this process of thinking and defining, we need to clarify that AI-generated works are also a form of intellectual labor. We must leave enough development space for technological progress and new creative methods while encouraging and protecting innovative human original works to benefit the development of cultural diversity.”

Literature is an important means of showcasing human spiritual originality. Will the large-scale application of AI weaken humans’ ability and opportunities to express their creativity?

Suo Luo bluntly pointed out the issues brought by AI writing entering online literature: “AI writing leads to the generalization of literary creation subjects. AI does not possess human subjective emotions; all the data resources it invokes are human creations. For literary creation that truly seeks to express subjective emotions and deep thoughts, the value of AI-assisted writing is minimal. This is why its impact on traditional literature is relatively small. The development of AI writing reflects the struggle between humanistic values and commercial interests in the literary field. It is necessary to defend human values while utilizing new technologies.”

The Corpus Training of AI Agents and Copyright Ownership

AI agents typically require a large amount of data for training, including publicly available texts, images, audio, and video. Whether the corpus used by intelligent platforms is obtained through legal means and whether they have the corresponding usage rights is a critical question. In recent years, copyright cases related to AI-generated content have involved novels, images, sounds, and other fields. How to determine the copyright ownership of AI-generated works? These issues not only concern the protection of creators’ rights but also directly affect the positive development of AI technology.

After the controversy over the “AI training supplementary agreement” on the Tomato Novel platform sparked resistance from contracted writers, Wang Feng keenly recognized the copyright issues brought by the new model of AI agent creation. He and his team members discussed the challenges faced by large model research and the writing field in depth. Regarding the legality of the corpus used by the platform, Wang Feng did not shy away: “Every AI agent faces huge challenges in this regard. The ‘Lingka Lingka’ creative writing AI agent does not currently involve corpus pre-training, but will undoubtedly face this difficulty later. We will obtain the corresponding training rights for the corpus through various legal means to protect the legitimate rights of original creators.”

The legality of the usage rights for the corpus used to train AI agents is not only related to the platform’s legal operation but also involves the protection of original authors’ rights and the balance of the entire content creation ecosystem.

In February 2024, the Guangzhou Internet Court ruled that an AI-generated Ultraman infringed copyright, stating that the plaintiff enjoys copyright over the Ultraman series image, and that an AI company generated images that were substantially similar to the plaintiff’s Ultraman image without permission, constituting reproduction and adaptation of the Ultraman image, thereby infringing copyright.

Artificial intelligence is a strategic technology leading the future, and China’s AI technology is also in a rapid development stage, making it equally important to balance the development of the AI industry and the protection of creators’ rights.

Xing Zheng believes: “The legal source of training data is a key issue that needs to be resolved in the development of AI writing. Both the training phase and the output phase of AI are equally important. AI works’ infringement is not only about infringement in the output phase but also includes infringement in the training phase.” Existing artificial intelligence training has a strong dependence on data, requiring a large amount of training samples for learning. From the perspective of intellectual property, original authors have not explicitly authorized their works to be used for AI training; unauthorized use may infringe the rights of original authors.

Do human users who collaborate with intelligent platforms to produce artworks own copyright? Since the overall artistic creation process is not entirely completed independently by users, they can only hold a portion of the share in the works. Regarding this “thorny issue,” Wang Feng outlined constructive solutions: dividing copyright into two steps, one for the product and one for the declaration and empowerment of copyright. Wang Feng believes that the future relationship between intelligent writing platforms and users will be that of collaborative creators, “In the final product, the platform and the user jointly hold the copyright and usage rights of the work. This is a new collaborative model, and of course, the usage of the final product can be negotiated based on agreements.”

How to resolve the issue of copyright ownership of AI-generated works? This requires a forward-looking intellectual property legal system that affirms copyright protection for generative artificial intelligence “works” and establishes a layered protection mechanism for human originals and works involving artificial intelligence participation in creation. Xing Zheng analyzed from a legal perspective that our current intellectual property legal protection system must take into account the particularity of AI creation. On one hand, it is necessary to strengthen legal research, improve relevant legal regulations, and clarify the copyright protection standards and rights attribution principles for AI works. On the other hand, it is also necessary to strengthen intellectual property protection through technological innovation, studying, protecting, and managing the AI agent creation model through technological means to prevent infringement. At the same time, it is essential to enhance industry self-discipline and regulatory efforts, promoting the establishment of industry standards and norms to provide strong guarantees for the healthy development of AI technology.

—END—

Source: Wenhui Daily

Author: Wang Xueying

Editor: Yang Wei

How AI Agents Address Copyright Concerns

Leave a Comment