It is difficult to distinguish between true and false, causing controversy. What is the applicable boundary of AI voice cloning?

Recently, simulated AI creations such as Stefanie Sun and Jay Chou have attracted widespread attention on the Internet and quickly aroused heated discussions among netizens. Due to its realistic nature, AI voices can not only reproduce the works of famous singers, but also raise a series of ethical and legal issues, and are even used by criminals, becoming an important means to increase the success rate of telecommunications fraud.

According to Science and Technology Daily, in telecom fraud, scammers use AI software to scan voice records and copy them, with an accuracy rate of up to 99%. Its corpus data is extracted from social media videos and other uploaded audio clips. After being input into the AI ​​software, it can generate voices and speak different phrases or sentences. According to foreign scholars, there are some models and algorithms that can generate a fairly high-quality cloned sound in just 1 minute or less. If more advanced artificial intelligence models and algorithms are used, the sounds produced will be more difficult for ordinary people to distinguish between true and false.

What is even more worrying is that when black and gray products such as AI voice cloning, AI face-changing, and personal information leakage are fully integrated, scammers can not only form electronic fraud "combination punches" through synthesized AI videos and voices, but also use Mastered private information and highly mature deception scripts generate deeply customized dialogue texts, escalating electronic fraud into a "parallel world" that is difficult to prevent, difficult to detect, difficult to wake up to, and nowhere to check.

In addition, the use of AI cloned audio in artistic fields such as singer-songwriters, songs, and music is also a strong challenge to the copyright field. The production of secondary creations may constitute infringement of the original author. At the same time, the copyright ownership of the work is also unclear and difficult to determine whether it belongs to humans or AI.

Just like the process that keeps repeating in history, technology comes first and institutions follow. During this blank period of norms and order, there is not only the conventional "barbaric growth", but also the unscrupulous tactics on the dark side. Tradition will also be broken by progress. and replace.

The above is the detailed content of It is difficult to distinguish between true and false, causing controversy. What is the applicable boundary of AI voice cloning?. For more information, please follow other related articles on the PHP Chinese website!