HearMe
As an AI-pilled enthusiast turned a worried declerationist, I made this album while interacting with the app HearMe. The app uses generative AI to support human-centered songwriting, encouraging users to play instruments, sing, write lyrics, and generate music in physical means. It also provides musical visualization aimed at explaining and teaching musical skills and concepts. The motivation behind HearMe is to address gaps in current musical AI generation, which pulls away humans from active and meaningful engagement with music making, rather than support and encourage the cognitive, emotional, physical and social benefits of physical music making. The music in the album was all played and sang by humans, after interacting and getting inspired by ideas generated by HearMe. This is because music should be played not promoted!
Q&A
01
Why the name HearMe?
The name reflects the human need to be heard in a world that seems to move towards AI-driven automation. It also refers to a proposal I wrote to fund the project titled “Human Ethics for Advancing Robotic Musical Intelligence” (HEAR MI).
02
What is the genre and mood of the album?
The album is a mix of indie rock and alternative electronic tracks. The mood is somber driven by anxieties about generative AI replacing meaningful human agency and preventing explainability in music making. There is also an unwavering thread of hope throughout the album.
03
What are the songs about?
The album has an overlying theme that addresses a wide variety of worries, from as mental illness (“Crown of Stars”) and aging (“When 60 Comes”) through political turmoil (“Storm”, “We’re Done”) and climate change (“Autumn Wind”), to concerns about AI (“HearMe)”.
04
How was the app HearMe used in making the album?
When working on a song, I would come up with lyrics and a few musical ideas often played on a piano, guitar, bass, or drums into HearMe. I would then listen to how the app puts these ideas together into a song and will use ideas from the AI generated composition to write my own song. Here is a short demo of an interaction with an early version of the app.
05
Can you describe specific song examples for how HearMe was used?
One example is the song "One More Things," where I left the original piano improvisation played into HearMe at the beginning of the song, before transitioning into a beat driven pieces that was composed based on ideas suggested by the app. Another example is the song “Everybody’s Talkin(g),” where I uploaded an audio sample of a failed effort to generate human speech into the app along with real-time MIDI piano and synth phrases I played. The app provided variations on these ideas that were then put together to make the song.
06
How is making music with HearMe different from how have made music in the past?
I am a jazz musician by training. Most of my music so far has been in that genre. During the pandemic, my students and I developed a generative AI app to write lyrics to some of my jazz tunes, which lead to the album "Better Late." In HearMe I reversed the process - it is the first time my research has inspired me to write my own lyrics, and to collaborate with AI in composing music for them.
07
How long did it take to make the album?
The first couple of songs in the album were written late 2024 / early 2025 when I was working with my band Haafarsekim. Back then, HearMe was just an idea, and for these first couple of songs, I used older AI tools I developed for inspiration. When starting to develop HearMe in 2025, I specified its features based on needs and gaps I identified in my previous research. The last song in the album “One More Thing” was composed in July 2026.
08
Who are the musicians on the album?
I played piano, keys, guitar, bass, drums and voice in all songs except:
-
Sharon Solly-Weinberg: voice on "Storm", "We're Done", and "Autumn Wind."
-
Guy Tesler: bass on "We're Done."
​