K-Pop Conference Finds AI Usage in Music Outpacing Copyright Law
... Korea is now the seventh largest music market in the world, according to the latest IFPI Global Music Report. It grew by more than 14% in 2022, significantly higher than the global average of 9.0%.
But the country’s copyright laws currently only cover humans and business entities, not virtual humans nor AI-generated content. And Korean courts have still not delivered their verdict on a patent application for Device for the Autonomous Bootstrapping of Unified Sentience (DABUS), an artificial intelligence system created by American scientist Stephen Thaler. DABUS’s patent applications have previously been rejected in countries including the U.S., UK, Australia and parts of Europe.
“As AI advances, we need to come up with synchronized international laws, so K-pop can spread throughout the whole world and IP can be protected,” said Kim Shi Hyeong, director general of the Korean Intellectual Property Office. ...
See the full story here: https://variety.com/2023/music/news/k-pop-music-conference-ai-artificial-intelligence-1235596468/
An Overview of Generative AI in Audio
| Generative Capability | Model Type(s) | Companies & Tools | Prompt / Input | Description | Use Cases |
|---|---|---|---|---|---|
| Song Lyrics | Text-to-text | ChatGPT(OpenAI), Moises, WaveAI | Natural language | Suggest song lyrics, including rhyming words and lines that follow a specific topic, structure (syllable count) and mood | Song lyrics composition |
| Music & Song Composition | Audio-to-audio | Jukebox and Musenet (OpenAI), Dance Diffusion (Harmonai), Amper (Shutterstock), DeepComposer (Amazon), Boomy, Mubert, AIVA, Endel, Amadeus Code | Tags/selected presets, natural language, and/or audio | Create, blend, autocomplete or refine music samples of novel melodies, instrumentation or other sonic elements, or adaptive soundscapes, customizable for specific moods, genre styles, or contexts | Music, songs, soundtracks, scores or themes for audio streaming, film, TV, games, metaverse spaces, podcasts, social video or ads |
| Deepfake Audio | Text-to-speech, speech-to-speech | Respeecher, Deepdub, Papercup, Resemble Al, MARVEL.ai (Veritone), VALL-E (Microsoft), Amazon Polly, Lovo, Eleven Labs, Replica Studios,VoiceMod | Speech and/or video | Create novel and naturalistic synthetic voices (singing or speech); replicate (clone) the speech profile and pattern of a specific voice; modify or augment tone, timbre, pitch or emotional quality; edit narration (adjust diction or delivery, redact or insert words) | Voiceover narration for film, TV, audiobooks, podcasts, news; voice dubbing in multiple languages for film, TV or other content that produces authentic accents while replicating the actor’s original voice tone and expressivity or adding expressivity not present in the original |
FOR MORE DATA, VISIT SOURCE: VIP+ ANALYSIS
Pages
- About Philip Lelyveld
- Mark and Addie Lelyveld Biographies
- Presentations and articles
- Tufts Alumni Bio