news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Our reporter Wei Zhezhe

Core Reading

AI speech synthesis, face recognition, voice interaction... my country's artificial intelligence industry is developing rapidly, and the scale of the industry continues to expand. While new technologies bring a better life experience, they also create new problems and disputes. Focusing on new formats and new developments, how to clarify boundaries and provide good protection in law, the reporter conducted interviews around typical cases.

Are sounds enhanced by technology protected? If I install a video doorbell at home, is there any basis for my neighbor to object? In recent years, the people's courts have fully played their functional role, properly resolved relevant disputes in accordance with the law, and used judicial judgments to regulate and safeguard the development of new business forms and new technologies, so that the people can better embrace new technologies.

Identify illegal activities and protect the legitimate personal rights and interests of AI voice and virtual images

With the widespread application of AI speech synthesis technology, as long as enough voice samples of a person are extracted, the voice can be "cloned" to produce related products. Technological updates make it easier to collect, synthesize, produce, imitate, and even tamper with voices, posing a challenge to the protection of voice rights.

Ms. Yin, a voice artist, encountered the problem of infringement of her voice rights. She found that some videos posted by users on short video platforms used AI dubbing based on her voice. After screening and tracing the source of the voices, Ms. Yin found that the voices in the above works came from a text-to-speech product on a platform operated by a Beijing smart technology company. Users can convert text into speech by inputting text and adjusting parameters.

It turned out that Ms. Yin had worked with a cultural media company to record an audiobook. Afterwards, the cultural media company provided the audio to a software company, which used the recordings Ms. Yin had made as materials for AI processing, generated a text-to-speech product, and sold it. After purchasing this product, a Beijing-based smart technology company directly retrieved and generated a text-to-speech product for sale on its platform without any technical processing.

Ms. Yin believed that her voice rights had been infringed, so she took the cultural media company, software company, etc. to court, demanding that they stop the infringement, apologize, and compensate for her economic losses.

According to the Civil Code, the protection of natural persons' voices shall refer to the relevant provisions on the protection of portrait rights. So, are voices enhanced by AI protected? "Voice rights are part of personal rights, and the premise for protection is that they are identifiable. Voices synthesized by artificial intelligence should be protected if the public can associate them with the natural person based on their timbre, intonation and pronunciation style," said Zhao Ruigang, vice president of the Beijing Internet Court. In the end, the Beijing Internet Court ruled that a Beijing smart technology company and software company should apologize to the plaintiff, and a cultural media company and software company should compensate the plaintiff for losses of 250,000 yuan.

As technology applications and business models continue to develop, technology is more deeply involved in the process of content creation and provision, and the boundaries between technology services and content services are becoming increasingly blurred. Some people think that "technological neutrality" means they can avoid responsibility, but this is not the case.

In a mobile accounting software, users can create their own "AI companions", set the companion's name, avatar and the relationship with the companion, and use common corpus to achieve communication and interaction. Public figure He found that he was set as a companion by a large number of users in the software. The company that developed and operated the software used a clustering algorithm to classify the companion "He" by identity, and promoted the virtual character to other users with a collaborative recommendation algorithm.

In this process, users uploaded a large number of portraits of the plaintiff as avatars. In order to make the virtual characters more humanized, the company also provided a "training" algorithm mechanism. In addition to the general corpus, users uploaded various interactive corpora such as texts, portraits, etc. that matched the personality of the virtual companion. The company used artificial intelligence to screen and classify them to form character-specific corpora, which were used in the conversations between the AI ​​companion "He" and users based on topic categories, personality characteristics, etc.

"The company does not provide a simple 'channel' service, but organizes users to create infringing materials and provide them to users through rule setting and algorithm design. The company's product design and the application of algorithms actually encourage and organize users' uploading behavior, which directly determines the realization of the core functions of the software. The company should bear the infringement liability as a content service provider, and the principle of 'technological neutrality' does not apply." Sun Mingxi, a full-time member of the Review Committee of the Beijing Internet Court, introduced.

The Beijing Internet Court held that the name, portrait, personality traits and other personality elements contained in the "virtual image" of a natural person are the objects of the natural person's personality rights. The creation and use of a natural person's virtual image without permission constitutes an infringement of the natural person's personality rights. In the end, the court ordered the company to apologize to He and compensate for economic losses, reasonable expenses and mental damages of 203,000 yuan.

Standardize the application scenarios of technologies such as face recognition and demonstrate a humanistic stance

Shao, who lives in a residential complex in Shanghai, installed a video doorbell on his entrance door that uses facial recognition technology and can automatically shoot and store videos. This practice caused dissatisfaction among neighbors in the adjacent building in the same community. Neighbor Huang said that the closest distance between the two homes was less than 20 meters, and the video doorbell was located directly opposite his bedroom and balcony. Based on the security monitoring facilities already in the community, Shao's behavior violated his privacy, so he asked Shao to remove the video doorbell. The two parties had a dispute over this and took the case to court.

"The Civil Code of my country stipulates that natural persons enjoy the right to privacy. No organization or individual may infringe upon the privacy of others by means of spying, intrusion, disclosure, or publicity." The trial court held that the privacy of a residence is the starting point and foundation for a peaceful personal life. Although Shao installed a video doorbell in his own space, the shooting range exceeded his own area, infringing Huang's privacy. In the end, the court ruled in favor of Huang's request to remove the video doorbell.

"This case explores the order of rights protection when the use of artificial intelligence devices conflicts with the enjoyment of privacy rights." A relevant person in charge of the First Civil Tribunal of the Supreme People's Court said that in order to properly and normatively use smart home products and avoid infringement of personal rights and interests, when the use of artificial intelligence devices conflicts with the enjoyment of privacy rights and personal information rights, attention should be paid to the priority protection of privacy rights and personal information rights and interests, highlighting a humanistic stance.

Face-scanning payment, face-scanning access control...Facial recognition technology has gradually penetrated into many aspects of life. While bringing convenience, it has also caused troubles for many people.

When Wang Moumou entered Guiyang East Station to take a train, the station broadcast reminded passengers that they needed to hold their ID cards and scan their faces to enter the station. Afterwards, Wang Moumou entered the station and took the train after scanning his face through the self-service ticket inspection channel. However, Wang Moumou believed that the collection of his facial information by China Railway Chengdu Bureau Group Co., Ltd. violated his legal rights and interests, so he filed a lawsuit in court, demanding that Chengdu Railway Bureau stop illegally collecting facial information and compensate for losses.

This case is the first public transportation facial recognition infringement dispute case in China since the implementation of the Personal Information Protection Law. How to balance public safety and personal information protection?

The Chengdu Railway Transportation Intermediate People's Court held that the railway department, based on its legal obligation to maintain public safety, processes passengers' facial information in accordance with the Personal Information Protection Law and does not require the passenger's personal consent.

Ultimately, the court took into account factors such as the Chengdu Railway Bureau providing passengers with the option of manual channels, multiple advertising notices, not excessive use of facial information, and the small impact and damage on Wang caused by the defect in the obligation to inform. The court concluded that the defect in the obligation to inform was not sufficient to constitute an infringement on its own, and the court did not support Wang's lawsuit.

Standardize the identification of interactive voice recognition and protect the rights and interests of technology-innovative enterprises

As a relatively mature human-computer interaction method, voice interaction is being widely used. Using a specific "wake-up word" for voice wake-up is one of the main wake-up methods for users to interact with smart devices.

In July 2017, a technology company released the first artificial intelligence speaker with the wake-up word "Xiao Ai", and later installed artificial intelligence voice interaction engines using the wake-up word "Xiao Ai" in mobile phones, TVs and other products. However, the company discovered that the wake-up word had been registered as a trademark by someone else.

From August 2017 to June 2020, Chen applied to register 66 trademarks including "Xiao Ai Tongxue" in different categories of goods. Later, he sent a lawyer's letter to a technology company's affiliated enterprises, demanding that they stop infringing on its "Xiao Ai Tongxue" trademark rights. He also used the "Xiao Ai Tongxue" trademark on sports watches, alarm clocks and other products with Shenzhen Yun Technology Co., Ltd. and jointly published product promotion articles. A technology company believed that the actions of Chen and Shenzhen Yun Technology Co., Ltd. constituted unfair competition and brought the case to court.

"After extensive publicity and use, 'Xiao Ai' can be used as a wake-up word with a certain influence, the name of an artificial intelligence voice interaction engine, and the name of products such as smart speakers equipped with artificial intelligence voice interaction engines, and is protected by the Anti-Unfair Competition Law." said Ye Tingzhou, a judge at the Wenzhou Intermediate People's Court in Zhejiang Province.

After hearing the case, the Wenzhou Intermediate People's Court held that Chen's actions of registering a large number of trademarks and sending "stop infringement" lawyers' letters violated the principle of honesty and trustworthiness, disrupted the fair market competition order, and also damaged the legitimate rights and interests of a certain technology company. They were unfair competition behaviors regulated by the Anti-Unfair Competition Law and constituted unfair competition by confusion and false advertising.

Ultimately, the Wenzhou Intermediate People's Court ordered that the infringement be stopped immediately, and Chen was ordered to compensate a certain technology company for economic losses and reasonable expenses of 1.2 million yuan, of which Shenzhen Yun Technology Co., Ltd. was jointly and severally liable for 250,000 yuan.

"The judgment in this case not only clearly protects the legitimate rights and interests of operators, but also effectively regulates the malicious registration of other people's wake-up words and abuse of rights, fully protects the brand reputation of technology-based innovative enterprises, responds to the operators' expectations for rights protection, and is of great significance for regulating business operations under new business formats and promoting fair competition in the market." said Wu Peicheng, a researcher at the School of Law of Zhejiang University.

People's Daily (Page 07, July 24, 2024)