一个月前一位朋友转发了个视频,对我说数字化转型网szhzxw.cn

真以为是孙燕姿本人翻唱了。前几天再次打开B站,密密麻麻的AI孙燕姿作品已经在我的首页霸屏。前奏一出瞬间把我拉回了,用随身听听磁带的岁月,人声出来后也毫不意外,很有孙燕姿的音色特色,除了唱腔和技巧差一点,人声和BGM的融合度很高,完全就是一首高分新歌。数字化转型网szhzxw.cn

一、“AI孙燕姿”翻唱周杰伦“出圈”
“AI孙燕姿”是否侵犯了孙燕姿的声音版权?是否侵犯了歌曲(例如周杰伦等歌手的歌曲)的著作权?未来AI会抢了歌手饭碗吗?数字化转型网szhzxw.cn
从AI换脸、AI绘画到AI写论文,科技的力量让人惊叹,现在AI翻唱又火出圈!最近,“AI孙燕姿”翻唱周杰伦的《发如雪》、南拳妈妈的《下雨天》在B站点击量均破百万,几乎就没有“AI孙燕姿”驾驭不了的曲风。目前网络上还活跃着“AI周杰伦”,“AI王心凌”等。那么问题来了,AI翻唱是否侵犯了歌手的声音版权?数字化转型网szhzxw.cn
二、“AI孙燕姿”火了 模拟音色还原度高数字化转型网szhzxw.cn
孙燕姿因其独具辨识度的音色与唱腔,成为AI宠儿,每天都有数十首“新作”面世。大家耳熟能详的《一直很安静》《岁月神偷》《红豆》,甚至《好汉歌》《向天再借五百年》等歌曲都在“AI孙燕姿”的演绎下有了别样的味道。扬子晚报/紫牛新闻记者随机听了几首后的感受是,声音非常逼真,仿佛孙燕姿真的亲自翻唱了这么多歌。数字化转型网szhzxw.cn
在现实生活中,孙燕姿如果想要公开演唱其他歌手的歌,无疑需要获得版权方的授权。目前,平台方和创作者主要通过声明AI作品不作商用来免责。
业内人士认为,按照目前的技术,歌手的唱腔、技巧和风格还很难完全模仿,但音色已经基本能1:1复制了。有一些歌迷,自己训练了已故歌手的AI,包括阿桑、张国荣、姚贝娜、邓丽君等,这或许是一种“数字永生”。数字化转型网szhzxw.cn
三、专家提醒:涉嫌侵犯声音权、著作权
河南泽槿律师事务所主任付建律师表示,对于音色识别度较高的歌手,声音就可以成为识别个人身份的重要依据,声音还具备一定的财产属性,在没有取得歌手授权的情况下,使用他人的声音、音色的,属于侵犯了歌手的声音权。数字化转型网szhzxw.cn
在付建看来,“AI孙燕姿”的核心技术是通过模拟歌手音色,使演唱出的作品近似原生。平台和创作者在互联网平台发布的作品未取得著作权人的授权,而且可能导致侵权行为扩大的风险,一份简单的免责声明并不能完全阻止侵权行为的发生,不能成为免责的事由。
“AI歌手翻唱涉嫌侵犯声音权、著作权等风险。”江苏苏延律师事务所丁律师表示,《民法典》第一千零二十三条规定,对姓名、自然人声音的保护参照适用肖像许可使用的有关规定。因此,未经歌手许可,使用AI技术将歌手的声音合成音乐侵害了声音权。另外,音乐作品的词曲均是有原创著作权人,著作权中包括发行权、复制权、表演权、信息网络传播权、改编权等,如未经著作权人许可使用软件合成或改编音乐作品让AI人物通过视频公开播放、表演,也侵害了他人的著作权。数字化转型网szhzxw.cn
四、未来,AI会不会抢了歌手的饭碗?
继AI绘图走红后,很多相关职业的打工人产生了焦虑,现在轮到歌手了吗?深度科技研究院院长张孝荣预测,未来AI有可能取代歌手,但需要在严格监管下才能实现。AI取代歌手的技术将给人类社会带来新威胁:深度伪造。通过深度伪造技术,个人形象和声音生物特征被AI复制操作,很容易产生新型的AI诈骗等更多社会问题。数字化转型网szhzxw.cn
付建律师也提醒,由于歌手的声音具有财产属性,可以轻易地成为他人谋利的工具,这也给了不法分子可乘之机。除了声音以外,歌手本人也是受到观众追捧的一个重要因素,AI脱离歌手本人后能否完全取代歌手,仍然取决于观众的审美和喜好。数字化转型网szhzxw.cn
文由数字化转型网(www.szhzxw.cn)转载而成,来源:网络;编辑:数字化转型网默然。

免责声明: 本网站(http://www.szhzxw.cn/)内容主要来自原创、合作媒体供稿和第三方投稿,凡在本网站出现的信息,均仅供参考。本网站将尽力确保所提供信息的准确性及可靠性,但不保证有关资料的准确性及可靠性,读者在使用前请进一步核实,并对任何自主决定的行为负责。本网站对有关资料所引致的错误、不确或遗漏,概不负任何法律责任。
本网站刊载的所有内容(包括但不仅限文字、图片、LOGO、音频、视频、软件、程序等) 版权归原作者所有。任何单位或个人认为本网站中的内容可能涉嫌侵犯其知识产权或存在不实内容时,请及时通知本站,予以删除。
翻译:
Video, said to me digital transformation network szhzxw.cn
I really thought Stefanie Sun sang it herself. A few days ago, I opened B station again, and the AI Stefanie Sun’s works have been dominating my home page. The prelude immediately pulled me back to the days when I listened to the tape with walkman. It was no surprise that the voice came out with Stefanie Sun’s timbres characteristics. In addition to the singing and skills, the voice and BGM were very high integration, which was completely a new song with high scores. Digital transformation network szhzxw.cn
I. “AI Stefanie Sun” cover of Jay Chou’s “Out of the Loop”
Does “AI Stefanie Sun” violate the copyright of Stefanie Sun’s voice? Does it infringe the copyright of a song (such as Jay Chou’s songs)? Will AI take Singer’s job in the future? Digital transformation network szhzxw.cn
From AI face changing, AI painting to AI writing papers, the power of technology is amazing, and now AI cover songs are becoming popular again! Recently, “AI Stefanie Sun” cover Jay Chou’s “Hair like snow” and Nan Quanma’s “Rainy Day” have been clicked more than one million times on B station. There is almost no music style that “AI Stefanie Sun” can’t handle. At present, the network is still active “AI Jay”, “AI Wang Xinling” and so on. So the question is, does the AI cover infringe the copyright of the singer’s voice? Digital transformation network szhzxw.cn
Two, “AI Stefanie Sun” fire analog tone restoration degree high digital transformation network szhzxw.cn
Stefanie Sun has become an AI favorite due to her distinctive voice and singing voice, with dozens of “new” songs released every day. Everyone familiar “has been very quiet” “Echoes of the Rainbow” “Red bean”, even “hero song” “to Heaven to borrow five hundred years” and other songs are in the “AI Stefanie Sun” interpretation has a different flavor. After listening to a few songs at random, the reporter of Yangtse Evening Post/Ziniu News felt that the sound was very realistic, as if Stefanie Sun had really covered so many songs herself. Digital transformation network szhzxw.cn
In real life, Stefanie Sun would no doubt need permission from the copyright owner if she wanted to perform other artists’ songs in public. Right now, platforms and creators are mostly exempt by stating that AI works are not commercial.
Insiders believe that according to the current technology, it is difficult to completely imitate the singer’s singing voice, skills and style, but the timbre has been basically 1:1 copy. Some fans have trained the late singers’ AI by themselves, including Ang Sang, Leslie Cheung, Yao Bena Na and Teresa Teng, which may be a kind of “digital immortality”. Digital transformation network szhzxw.cn
Three, experts remind: suspected infringement of voice rights, copyright
Fu Jian, director of Henan Zejin Law Firm, said that for singers with high timbre recognition, voice can become an important basis for personal identity identification, voice also has certain property properties, without obtaining the singer’s authorization, the use of others’ voice, timbre is a violation of the singer’s voice rights. Digital transformation network szhzxw.cn
According to Fu Jian, the core technology of “AI Stefanie Sun” is to simulate the singer’s voice and make her songs resemble the original ones. Works released by platforms and creators on Internet platforms have not obtained the authorization of copyright owners, and may lead to the risk of expanding infringement acts. A simple disclaimer cannot completely prevent the occurrence of infringement acts, and cannot be the cause of exemption.
“AI singers are suspected of infringing voice rights and Copyrights.” Ding, a lawyer with Jiangsu Suyan Law Firm, said that Article 1023 of the Civil Code stipulates that the protection of names and natural persons’ voices should refer to the relevant provisions applicable to the licensing of portraits. Therefore, the use of AI technology to synthesize music from a singer’s voice without the singer’s permission violates the right to sound. In addition, lyrics and songs of music works are all owned by original copyright owners, including the right of distribution, reproduction, performance, information network communication, adaptation, etc. If music works are synthesized or adapted by software for AI characters to play and perform publicly through videos without the permission of copyright owners, the copyright of others is also infringed. Digital transformation network szhzxw.cn
Fourth, in the future, will AI take the singer’s job?
After the popularity of AI graphics, many workers in related professions have been anxious. Now is it singers’ turn? Zhang Xiaorong, head of the Deepin Institute of Science and Technology, predicted that AI might replace singers in the future, but only under strict supervision. AI replacing singers poses a new threat to human society: deep counterfeiting. Through deep forging technology, personal image and voice biometrics are copied and manipulated by AI, which easily leads to new types of AI fraud and other social problems. Digital transformation network szhzxw.cn
Fu also warned that since a singer’s voice has property properties, it can easily be used as a tool for others’ profit, which also gives criminals opportunities. In addition to the voice, the singer himself is also an important factor to be pursued by the audience. Whether AI can completely replace the singer after being separated from the singer himself still depends on the audience’s aesthetic taste and preference.
