Why Splice's CEO Says AI Will Never Replace Real Musicians
Splice CEO Kakul Srivastava says AI will never replace human creativity, and she wants Splice to be built around that line.
She explained her thinking in interviews with The Verge, including a Decoder talk with Nilay Patel published March 24, 2025 about where to set firm limits for AI in music, and a follow-up published Oct. 3, 2026. The Verge
Srivastava came to Splice as CEO from Adobe after sitting on its board since 2021. Founder Steve Martocci then became Executive Chairman and Chief Strategy Officer. She had been CEO for a little over six months as of December 2022. She had earlier worked at Flickr, Yahoo, GitHub and Adobe. The Verge She studied Mechanical Engineering at MIT and earned an MBA from UC Berkeley's Haas School of Business. She was listed among TIME's 100 Most Influential People in AI in 2025.
Splice is known for its library of samples and tools for makers. Its Sounds Plugin, now in testing, brings millions of royalty-free sounds that can be legally reused into a DAW, the main software musicians use to record and mix. Users can search, listen, drag and drop, and make variations inside a song project, like pulling ingredients from a well stocked pantry. The company bought Spitfire Audio. Splice samples were used in Lisa's "Money" and Sabrina Carpenter's "Espresso."
Splice is a remote-first company. Its research and education work includes Alejandro Koretzky, Vice President of Applied Research, and Meredith Allen, Director of Business Development, Education, who both spoke for Splice at the 2026 NAMM Show.
Her most important tool is not software but process. She described it as "a well-written document and a real conversation about it." She is careful with documents written by AI because "once you can't tell whether there is a person behind the words, the conversation becomes different and worse."
On lighter questions, she picked Instagram as her happy place online, Antakshari as her favorite childhood game, and a cup warmer as her most disappointing gadget.
The broader context here helps explain the stance. In my view, Srivastava is talking about two rules. One covers music itself, which sounds are okay to use and release. The other covers teamwork, which plans still show a real person made a choice. Fast search helps. Clear credit keeps trust.
In my view for people building tools, the question is not if AI can make music that sounds okay. It can. It is whether musicians, producers, teachers and rights holders will accept how it was made. A remote-first library needs trust, clear permission rules and tools that do not slow down a session. Keeping people visible in plans and performances protects that trust. That still leaves room for better search, better tools and better learning, without giving authorship to autocomplete.
