Nandu News Reporter Li Ling Recently, foreign media reported that the American technology giant Meta is negotiating with a number of Hollywood stars for AI voice projects, and plans to pay millions of dollars to obtain their recordings and the rights to use their voices in AI projects.
stars involved in the negotiations include Judi Dench, Awkwafina and Keegan-Michael Key. Meta plans to launch a series of AI tools at the connect conference in September this year. In order to gain more research and development time, meta is currently negotiating with these stars.
Sources revealed that meta hopes to use the voice data of these stars to train AI models, develop chat robots, and provide services such as digital assistants or virtual friends, allowing users to talk to robots with the voices of Hollywood stars.
The details of meta's negotiations with these stars are currently unclear. However, due to differences in terms of voice usage, negotiations between meta and some celebrity representatives were interrupted several times. According to reports, meta wants broad usage rights for a fixed period, while celebrity representatives are asking for tighter restrictions.
This time, meta actively seeks the voice authorization of Hollywood stars, undoubtedly to avoid infringement disputes. After all, a few months ago, openai was involved in the controversy.
In May this year, openai released the large model gpt-4o, claiming to have made a breakthrough in speech generation. This new version comes with a total of 5 voices, one of which is called sky, which is said to be "highly similar" to the famous movie star Scarlett Johansson.
Scarlett later posted that she had clearly rejected openai’s dubbing invitation in September last year and was shocked, angry and unbelievable that such a situation still occurred. In order to protect her own rights and interests, Scarlett sent a lawyer's letter to OpenAI, requesting an explanation of the exact creation process of Sky.
openai responded that the use of AI-synthesized speech was within the scope of fair use, but out of respect for Scarlett Johansson, it announced that it would suspend the use of sky speech. The incident of
has also triggered a lot of discussion. With the widespread application of AI speech synthesis technology, it is increasingly common for sounds to be collected, synthesized, produced, imitated and even tampered with. While AI technology is innovating, how to protect personal rights including voice, portrait and privacy rights?
Nandu reporters noticed that on August 7, the official WeChat account of the Supreme People’s Court pushed a case involving AI-generated sounds. The case was heard by the Beijing Internet Court, which clearly determined that the protection scope of natural persons’ voice rights can extend to AI-generated voices, provided that they are identifiable.
This is the country’s first “AI-generated voice personality infringement case”. The plaintiff in this case, Ms. Yin, is a voice actor. After discovering that her voice had been artificially altered, she claimed to the court that the five defendants had violated her voice rights.
Ms. Yin said in an interview with the media that she has never authorized any person or company to use AI for her voice, but some short movie commentary and game commentary videos use AI dubbing based on her voice.
It turned out that Ms. Yin had cooperated with a cultural media company and recorded several audiobooks. Afterwards, the company provided the audio recordings to relevant software companies for AI processing, and generated a text-to-speech product for sale. The defendant in this case, a smart technology company, purchased this product and sold it directly to external parties without any technical processing.
On April 23 this year, the Beijing Internet Court issued a first-instance verdict. The court held that the authorization of sound recordings does not mean the authorization of sound ai, and unauthorized use or permission of others to use the sounds in sound recordings without the permission of the rights holder constitutes infringement.
In the end, the court ruled that a certain smart technology company and a certain software company should apologize to the plaintiff, and a certain cultural media company and a certain software company should compensate the plaintiff for losses totaling 250,000 yuan. After the first-instance verdict was pronounced in this case, neither party filed an appeal, which is now in effect.
With the widespread application of AI synthesis technology, from Hollywood stars to dubbing artists to every ordinary user, in fact, they may face the trouble of being imitated or having their faces changed by AI. What should we do about this?
Scarlett wrote in the statement blaming openai: "As we all grapple with deepfakes and protecting our own likenesses, work, and identities, I believe these issues must be made clear. I look forward to more transparency approach to solve the problem and ensure the protection of individual rights through appropriate legislation.”
.