laitimes

Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement

author:Frontier of intellectual property
Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement
Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement

1. Case introduction

The plaintiff in this case, Yin, is a dubbing artist who accidentally discovered that his dubbing was used by others to produce various works and widely disseminated on multiple app platforms. After tracing the source, Yin found that his voice in the work came from the text-to-speech function of the platform operated by Defendant Company A, and as long as the text was entered, he could generate a voice consistent with his own voice. As a result, Yin filed a lawsuit against the five companies involved in the case.

Due to the complexity of the case and the involvement of five companies as defendants, the team of Sister Sa sorted out the relationship between the plaintiff and the five defendants into the following chart for easy understanding.

Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement

As can be seen from the above figure, the infringement in this case occurred because Company B, as the copyright owner of the recording product, provided the audio recording product to Company C without authorization and allowed Company C to use it commercially or non-commercially, resulting in Company C using the audio recording product containing Yin's voice provided by Company B to form an AI voice product, and Company C sold the product online on Company D's platform, which eventually led to Yin's discovery of the product involved in the case from Company A.

After the trial, the court finally ordered Company A and Company C to apologize to the plaintiff, and Company B and Company C to compensate the plaintiff for losses of 250,000 yuan.

2. Case analysis

(1) How should sound infringement be determined?

Paragraph 2 of Article 1023 of the Civil Code clearly stipulates that the protection of the voice of natural persons shall be governed by reference to the relevant provisions on the protection of portrait rights. Therefore, no one may use the voice of the owner of the sound, unless otherwise provided by law, without the consent of the owner of the sound. Under this premise, even if Company B is the copyright owner of the sound recording products containing Yin's voice, it does not have the right to use or license others to use Yin's voice unless authorized by Yin, so Company B's unauthorized provision of sound recordings to Company C for Company C's AI training to generate AI voice products certainly infringes Yin's relevant personality rights to his own voice.

In this case, it is worth noting that the court has made a more detailed discussion on whether the plaintiff's voice rights and interests extend to the AI voice in question. The reason for this is that if the rights and interests of the voice are not as good as the AI voice involved in the case, then except for the direct infringers Company B and Company C, the remaining defendants will not constitute an infringement of the voice rights and interests of the plaintiff Yin because they only use the AI voice involved in the case. In this regard, the court held that it should be considered from whether the AI voice involved in the case is recognizable and whether it can be associated with a specific natural person.

Specifically, the court pointed out that the voice of a natural person is distinguished by voiceprint, timbre and frequency, and has the characteristics of uniqueness, uniqueness and stability, which can form or cause ordinary people to have thoughts or emotional activities related to the natural person, and can show the individual's behavior and identity to the outside world. Therefore, if the AI voice involved in the case can enable the general public or the public in related fields to associate with a specific natural person based on timbre, intonation, style, etc., then the voice should be considered to be recognizable. Accordingly, the court determined that the AI voice involved in the case was recognizable, and the plaintiff's voice rights and interests were extended to the AI voice involved in the case.

Therefore, considering that Company A, Company D, and Company E all used the AI voice involved in the case, they need to bear certain tort liability on the basis of satisfying the constitutive elements of other infringements.

(2) How to confirm tort liability?

The defendants raised a defense against the tort liability in this case, but the court did not adopt all of them, but reasoned with the defendants' defenses and properly handled the issue of the defendants' liability, as follows:

Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement

Therefore, after comprehensively considering the circumstances of the infringement, the value of products in the same market, the number of broadcasts of the products and other factors, the court finally ruled that Company B and Company C should bear the liability for damages of 250,000 yuan.

The tort liability of the Civil Code of the People's Republic of China is based on the principle of fault liability, so whether the actor is subjectively at fault is the main constitutive element for judging tort liability. In this case, Company B, as the copyright owner of the sound recording, obviously knew that it had not obtained the plaintiff's authorization to use the sound, so it was naturally difficult to say that it was not subjectively at fault, and in the same way, Company C, as a direct collaborator with Company B, would be difficult to deny the subjective fault of Company C unless it requested Company B to provide proof of the relevant authorization. On this basis, as a platform service provider, Company D only needs to fulfill its general duty of care, and it is generally not liable unless the infringement of the sound is so obvious that Company D knew or should have known that the product was suspected of infringement. As for Company A and Company E, since one party was only entrusted by the other party to purchase the products of Company C, neither party participated in any of the aforesaid related acts, so it was difficult to know the fact of infringement, and there was naturally no subjective fault.

To sum up, the determination of tort liability in this case is extremely reasonable.

3. Write at the end

As mentioned above, the rollout of AI technology is bound to have a large number of infringement cases under the current legal framework, but how to determine infringement and how to accurately determine the existence and magnitude of infringement liability is still a matter that needs to be carefully considered.

Source: Xiao Sa lawyer

编辑:lancelot

Xiao Sa Team | The first case of AI-generated voice personality infringement!Analysis of the risk of voice infringement