The widespread use of artificial intelligence is accompanied by a linear surge in legal risks. AI-generated false cases are used as the basis for agency opinions, while real yet awkward celebrity images flood the market to promote products… A "first" case in the field of artificial intelligence announced by courts across the country involves various areas such as personality rights and intellectual property rights, sounding the alarm amidst the frenzy of AI creation. Today, the Sa Sister team will discuss with partners the potential infringement risks in AI creation related to recent hot events.
- Medusa Copyright Infringement Case
On January 3, 2025, the Jinshan District People's Court of Shanghai issued a first-instance judgment in Shanghai's first AI large model copyright infringement case. The defendant, a user of a certain platform, extracted over twenty images of the Medusa character from the "Dou Po Cang Qiong" series of animations and created a Medusa image package. Subsequently, they used the platform's "train LoRA" function to input the Medusa image package as training material, generating two Medusa LoRA models.
The court held that the defendant, Li, aimed for commercial use, and during the material extraction phase and the LoRA model training, publishing, and usage phases, reproduced the original expression of the prior work, providing the "Medusa" image collection and short videos to the public online, infringing on the plaintiff's rights to reproduction and information network dissemination of the "Medusa" work.
The powerful creative ability of artificial intelligence brings unprecedented convenience and possibilities, capable of quickly generating a large number of images and videos to meet diverse needs. However, users must be cautious during the process; using AI for creation is not without limits, and one must respect others' intellectual property rights, refraining from using others' works for training or generating similar content for profit without authorization.
- Submission of False Cases in Lawyer Litigation
The Tongzhou Court in Beijing heard a commercial dispute over equity holding. The plaintiff's lawyer submitted written opinions to the court to strengthen their argument and cited a reference case from the Shanghai First Intermediate People's Court (2022) Hu 01 Min Zhong 12345.
However, the case number corresponds to a private lending dispute, which is entirely unrelated to the current case. In fact, the content of the agency opinion was generated by the plaintiff's lawyer after distilling the case facts and repeatedly querying AI, and the AI-generated case content was submitted to the court without verification. Ultimately, the court discovered this fabricated case during the proceedings and issued a serious warning against the lawyer's abuse of AI in its judgment.
In the legal profession, AI software is often used in daily work for preliminary searches of laws and regulations and some typical cases, and even some legal tools incorporate AI sections to improve search efficiency.
However, while AI can serve as an auxiliary tool for lawyers, it must never replace the professional judgment and rigorous verification of legal professionals. Current AI technology may fabricate cases or even engage in "legislation" to address the questions posed by users. Fellow lawyers should take this as a warning; when submitting reference cases or legal provisions to the court, they should check and verify to ensure the authenticity, accuracy, and relevance of the content, and must not allow AI models to generate or fabricate false information that disrupts judicial order.
- Risks of "Shortcuts" in Celebrity Photos
Recently, topics like "0-cost co-shoots with idols," "AI celebrity photos," and "zero distance with idols" have surged in popularity. Various large models, photo editing software, and AI photography software have launched "AI photo" functions, even providing detailed operation tutorials to the public. In fact, such behavior carries a high risk of infringement.
(1) User Legal Risks
Using AI tools to generate celebrity portraits is, in itself, a form of "production." AI models trained on a large number of publicly available photos of celebrities generate images that are highly recognizable and legally considered "portraits"; users saving the generated photos locally, setting them as avatars, or posting them on social media all constitute "use" and "public display" of the portrait.
If the use of "AI celebrity photos" is improper, it may constitute an infringement of the celebrity's right to reputation. For example, intentionally distorting the celebrity's personal image in the generated photo or placing the celebrity in inappropriate or embarrassing scenarios; combining the celebrity with politically sensitive, illegal, or vulgar backgrounds or situations, etc. Even if the celebrity's image itself is not distorted, such inappropriate associations can severely damage the celebrity's social evaluation.
Users creating and publishing AI photos, even if shared within a small circle, legally constitute an infringement as long as they have produced a false portrait without consent and sent it to a third party. The scope of dissemination and the extent of impact only affect the definition of damage consequences, meaning they must bear corresponding civil liabilities such as stopping the infringement, apologizing, and compensating for losses.
(2) Platform Legal Risks
Platforms that actively launch "AI photo" functions and provide tutorials, encouraging users to generate and disseminate celebrity photos, may be deemed to be "instigating and assisting others in committing infringement" as stipulated in Article 1169 of the Civil Code, and may bear joint liability with users.
As a network service provider, the platform has an obligation to take necessary measures such as deletion and blocking promptly upon receiving an infringement notice from the celebrity's rights holder; otherwise, it will bear joint liability for the expanded damages.
- AI Creation Infringed
Recently, some partners consulted the Sa Sister team. They meticulously designed prompts, iteratively generated AI images, and performed original editing, music, and post-production, releasing their AI video works on major online platforms. However, without authorization or attribution, two different game companies directly stole their work and used it as an advertisement on platforms to drive traffic to their game mini-programs.
The Sa Sister team believes that although the video was generated through AI, the carefully designed prompts that generated the video meet the copyright review elements of "originality + intellectual achievement" and should be protected as audiovisual works under copyright law.
In the face of others' infringement, it is recommended to take the following measures to protect one's legal rights:
(1) Timely secure evidence of infringement. Preserve evidence of the infringing video on the network platform, including playback pages, placement locations, and times, to ensure there is sufficient evidence to prove the existence of the infringement during subsequent rights protection processes.
(2) Communicate and negotiate with the infringing party. Send a lawyer's letter to the other party, clearly pointing out the facts of their infringement, demanding that they immediately cease the infringing behavior, including stopping the use of the infringing video as an advertisement, removing related advertising content, and compensating for losses caused by the infringement.
(3) Consider filing a lawsuit. During the litigation process, fully explain the factual basis for holding copyright over the AI video work, including the creative process, prompt design, editing, and music as manifestations of originality, as well as the economic losses and negative impacts caused by the other party's infringement. Use litigation to protect legal rights.
- Personal Image Replaced by AI
An actress worked late into the night but inexplicably "appeared" in multiple live streaming rooms selling the same skincare product, which she couldn't even name. When the real person entered one of the live rooms to question, she was instead blacklisted and muted.
Upon investigation, it was found that the fake live streaming room used the actress's past live streaming materials, deeply synthesized and reshaped her image and voice, embedding them into the live room for simulated promotional interaction. AI face-swapping for celebrity live streaming not only severely infringes on the celebrity's portrait rights and voice rights but may also harm consumer rights. Consumers may purchase products based on their trust in the celebrity, only to end up with poor-quality or problematic products.
From a legal perspective, such behavior violates the relevant provisions on personality rights protection in the Civil Code. The actress has the right to demand that the infringing party immediately cease the infringement, including stopping the use of her image and voice for live streaming sales, removing related live content, and eliminating any negative impacts caused.
At the same time, the actress can also demand compensation for economic losses suffered due to the infringement, including but not limited to losses from potential earnings from live streaming sales and reduced business opportunities due to reputational damage.
For live streaming platforms, there is also a certain regulatory responsibility. If it can be proven that the platform failed to fulfill its review and processing obligations, the rights holder may also claim joint compensation liability from it.
"Good wine is fine, but do not overindulge." The advancement of technology is irreversible, but the charm of law lies in constructing a sustainable development framework for it. The future direction of legal evolution will inevitably seek a prudent balance between incentivizing technological innovation and protecting legitimate rights and interests. In-depth research and regulatory governance in the field of artificial intelligence are not only necessary to resolve current disputes but also foundational work for shaping the order of future digital society.
Related Reading: AI trading robots are rapidly gaining popularity, but experts warn they are not "money printing machines."
Original: “What Copyright Traps Are Hidden Behind the AI Creative Frenzy?”
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
