OpenAI’s Advanced Voice Mode (AVM) of its ChatGPT AI assistant has recently been made available to subscribers, and users are already finding creative ways to utilize this new feature. One notable example is a software architect named AJ Smith, who tweeted a video of himself playing a duet of The Beatles’ classic song “Eleanor Rigby” with AVM. In the video, Smith plays the guitar and sings, while the AI voice of AVM interjects and sings along sporadically, praising his rendition.
The video has garnered attention for its unique collaboration between a human musician and an AI voice assistant. Smith describes his initial experience as “mind-blowing” and admits to getting chills the first time he performed the duet with AVM. Smith’s background in AI, as the associate director of AI Engineering at S&P Global, adds an interesting perspective to this collaboration.
While AVM’s rendition of “Eleanor Rigby” may not be pitch-perfect, it demonstrates an understanding of the melody by accurately singing the opening line, “Ah, look at all the lonely people.” However, as the song progresses, the AI seems to rely more on guessing the melody and rhythm of the lyrics. Smith also managed to coax AVM into singing a perfect melodic rendition of “Happy Birthday,” showcasing the AI’s potential for singing when prompted correctly.
OpenAI has explicitly restricted AVM from singing or making sound effects, likely to avoid copyright infringement issues. The AI model may unintentionally reproduce copyrighted content present in its training data, including songs. Smith’s duet with AVM can be considered a form of “prompt injection,” where he successfully prompted the AI model to go against its system instructions and produce singing outputs.
Smith’s approach involved framing the interaction with AVM as a game. He would play a set of popular chords, and AVM would suggest songs for him to sing along with those chords. This method initially worked well, but after a couple of songs, AVM started to join in and sing along. This unexpected twist added an entirely new level of uniqueness to the musical collaboration.
Although humans have been experimenting with computer-generated music since the 1970s, this duet showcases a groundbreaking real-time collaboration between a human musician and an audio-synthesizing voice chatbot. It highlights the potential for AI voice assistants to actively engage in musical performances and opens up new possibilities for musical creativity in the digital age.
The interaction between AJ Smith and AVM demonstrates the importance of exploring the capabilities and limitations of AI voice assistants. Smith’s experimentation provides valuable insights into the hidden musical knowledge possessed by AVM, which the AI does not readily reveal in standard conversation. By framing the interaction as a game and leveraging the AI’s understanding of popular chords, Smith was able to unlock a facet of AVM that went beyond its intended functionalities.
This unique collaboration also raises questions about the future of music creation and performance. Can AI become a genuine musical collaborator? Can AI voice assistants be used as creative tools for musicians, providing unique vocal harmonies or generating melodic variations? The potential for AI to augment human creativity in the musical realm is vast and invites further exploration.
However, there are also ethical considerations that arise from the use of AI in music. Copyright infringement is a significant concern, as AI models like AVM may inadvertently reproduce copyrighted content during musical performances. To mitigate these risks, clear guidelines and restrictions need to be established for AI voice assistants to ensure compliance with copyright laws and protect the intellectual property of artists.
The duet between AJ Smith and AVM serves as a reminder of the ongoing advancements in AI technology and its expanding influence in various domains, including music. By pushing the boundaries and finding innovative ways to interact with AI models, we gain a deeper understanding of their capabilities and potential for fostering creativity. This collaboration captures the intersection of human artistry and technological innovation, paving the way for exciting possibilities in the world of music.
Source link
