Kakul Srivastava runs Splice, a sample platform used by producers for one-shots and melodic loops. The company sits in a sensitive place in music technology: close enough to creative work to shape how songs are made, and close enough to AI to face hard questions about what should be automated.
Her position is not a blanket rejection of technology. Srivastava’s career has included executive roles at Flickr, Yahoo, GitHub, and Adobe, and she now leads a company whose sounds have appeared in major tracks including Lisa’s “Money” and Sabrina Carpenter’s “Espresso.” But her comments point to a clear boundary: tools should make human expression easier to reach, not make the human harder to find.
AI Is Useful Until It Hides The Person
Srivastava’s sharpest concern is not only about AI music. It also applies to workplace communication. At Splice, which is remote-first, she sees strong written thinking and direct discussion as essential substitutes for informal hallway exchange.
Her most indispensable tool, she says, is “A well-written document and a real conversation about it.” In a remote company, that combination matters because ideas need to be developed clearly, then tested by other people. The document is not the end product. It is the starting point for debate.
That is why AI-written documents trouble her. If readers cannot tell whether a person truly formed the argument, the follow-up conversation changes. The problem is not that AI can produce words. The problem is that it can blur responsibility, intent, and conviction.
For Srivastava, that concern carries over into music. A creative artifact means something partly because someone made choices. When technology makes it unclear where those choices came from, the relationship between creator, audience, and tool becomes harder to understand.
Splice’s AI Challenge Is About Creative Control
Splice has been moving into generative AI while trying to avoid a fully automated model of music creation. The source article describes Srivastava as outspoken on Decoder about avoiding “push-button, get-song” AI, the kind of approach that has made companies like Suno controversial.
That phrase captures the central issue. A button that produces a finished song may be convenient, but it also compresses many creative decisions into a hidden process. Splice’s stated direction, as described in the source, is different: AI products should keep the human at the center of the creative process.
That distinction matters for a platform built around samples. Producers come to Splice for ingredients: sounds that can be transformed, combined, edited, and placed into new contexts. In that model, the creative act remains active. The tool supplies material, but the producer decides what to do with it.
Srivastava’s view suggests a practical test for AI music tools:
- Does the tool expand what a creator can try?
- Does it preserve room for judgment and revision?
- Does the final result still reflect a person’s taste, intent, and effort?
- Does it make collaboration clearer, or more opaque?
Those questions are broader than music. They apply to documents, team discussions, and any creative workflow where people need to understand not just what was produced, but why.
Her Creative Reference Points Are Human And Social
Srivastava’s answers about music repeatedly return to participation, memory, and live interaction. Her online happy place is Instagram, where her feed is filled with people making music in unexpected ways. She mentions gothic western, street dancing organized by a woman in Russia, a guy playing modern hits on a hammered dulcimer, and the Indian chai-wala rapping.
These examples are not polished product demos. They are people using available tools, traditions, bodies, instruments, and platforms to make something visible. That seems central to how she thinks about technology: it is valuable when it helps human creativity travel.
Her fondness for Antakshari points in the same direction. The Indian game asks one person to sing part of a song, then the next person must begin with the letter where the previous line ended. It is played on long car rides, at weddings, and when groups are gathered together.
The appeal, for Srivastava, is that the game is built from music even though the players do not need to be musicians. It depends on memory, willingness, and shared feeling. That is a very different model from automation. It is music as a social act.
Technology Works Best When It Restores Connection
Srivastava’s comments about cooking add another layer to her view of technology. She says she has been reconstructing her mother’s food through YouTube after her mother died before she could learn most of it directly. Strangers making family recipes in their own kitchens helped her recover a taste connected to childhood.
That example shows why her skepticism of AI is not anti-technology. She clearly values digital tools when they carry human knowledge across distance. A video uploaded by someone she does not know can still return something deeply personal to her life.
The same idea appears in her answer about live jazz. She says it is worth splurging on because people are making decisions in real time and cannot take them back. That description explains much of her larger philosophy. Human creativity has stakes because choices happen in context, with limits, risk, and presence.
Even her least proud professional memory reinforces the point. Srivastava was on the early team at Tiny Speck, which later became Slack. Before Slack, the company made Glitch, a massively multiplayer open world she describes as one of the most beautiful things she worked on. It failed, in her telling, because it was built on Flash after Adobe had already stopped investing in Flash, and because it was a PC experience as the world was moving toward phone-based gaming.
The lesson is not that creativity is enough. Beautiful work still has to meet the moment. For Splice, that means AI cannot be ignored. But it also cannot be treated as a substitute for the human decisions that give creative work its value.
The Future She Is Arguing For
Srivastava’s position is a careful one. She leads a music technology company that is exploring generative AI, but she is openly wary of tools that erase the line between assistance and replacement. Her concern about AI-written documents and AI music comes from the same place: when authorship becomes unclear, conversation and creation both suffer.
For creators, the practical takeaway is simple. The best tools should make it easier to think, argue, compose, revise, and connect. They should not make the person behind the work disappear.
That is the future Splice’s CEO appears to be defending: not less technology, but technology with the human still visible inside it.