Scientists from the United States have presented an instrument that allows you to play verses in the style of famous rappers. Researchers are ready to remove audio at the first request of the copyright holders.
If you’ve ever dreamed of creating a song with Tupac or Jay-Z, then Uberduck’s new artificial intelligence (AI) tool could bring you closer to that goal. Uberduck allows you to select the voice of a celebrity and then enter the text that she should say.
For example, the instrument can synthesize speech for Tupac’s “calm” or “tense” verse. Users can also play Eminem’s verses in different styles – for example, before and after The Eminem Show album.
The creator of Uberduck says they have begun work on a system with the goal of creating an open conversational agent.
The researchers note that the synthesized voices are still “far from perfect,” but through changes to the text – for example, adding extra vowels to lengthen the syllable – users can make very similar imitations of their favorite rappers.
Scientists understand that their tool can be used for misinformation and defamation. The developers have tried to minimize this risk – users are prohibited from using the audio for commercial purposes or to create defamatory material. They must also indicate that their artwork was created by artificial intelligence. Researchers are ready to remove audio at the first request of the copyright holder.
At the same time, the company notes that their activities are legal – in this case, we are talking about parodying, not recreating the voice.