Splice CEO Kakul Srivastava Warns Against Replacing Human Effort With AI
Splice CEO Kakul Srivastava explains why automated emails and documents damage workplace collaboration, emphasizing the need for human-centered technology.
As generative artificial intelligence spreads across workflow software, corporate communications, and digital media production, Splice Chief Executive Officer Kakul Srivastava is urging caution regarding automated text and audio generation. Srivastava, whose audio sampling platform provides material used in major hits such as Lisa's "Money" and Sabrina Carpenter's "Espresso," argues that displacing human effort in written communication and creative workflows alters conversations for the worse.
Srivastava has spent her career working at the intersection of technology and creative fields. Before guiding Splice through its acquisition of Spitfire Audio and its expansion into human-centered generative AI tools, she held key executive roles at Flickr, Yahoo, GitHub, Adobe, and Tiny Speck. Based on her experience, she remains skeptical of "push-button, get-song" systems like Suno that bypass human input entirely, applying those same concerns to workplace communication.
The risks of automated workplace communication
Splice operates as a remote-first organization without a physical hallway for spontaneous discussions. Instead, the company relies on employees sitting down to write thoroughly considered documents, which are then scrutinized and debated by colleagues. Srivastava notes that almost every successful initiative at the company originated from this specific process of written thought and debate.
However, the introduction of AI-written documents threatens this dynamic. Srivastava cautions that once team members can no longer tell if a real person is behind the text, the nature of the conversation changes and degrades. In her view, machine-written documents strip away the authenticity required for meaningful critique, creating a point of no return for team communication. She sees a direct equivalence between this breakdown in writing and the automated generation of music.
Her focus on human connection extends to how she views technology's ultimate purpose. Despite past industry failures—such as her time on the early team at Tiny Speck working on Glitch, an open-world game that failed due to relying on Flash and missing the transition from PC to mobile—Srivastava sees technology at its best when it serves human memory and expression. She uses YouTube videos uploaded by strangers to reconstruct her late mother's recipes, finding that digital platforms are most valuable when preserving real human experiences.
What it means for developers
For software engineers and product creators building AI-assisted tools, Srivastava's observations highlight the boundary between helpful automation and counterproductive clutter. When developing features for remote collaboration, documentation, or creative tasks, blindly injecting generative text can obscure genuine intent and weaken critical feedback loops.
Engineers building software for creative or collaborative environments need to decide whether an AI feature aids human judgment or merely generates noise. Designing systems that keep the human user at the center requires careful prompt engineering, UI design, and model evaluation. To explore these workflows across different providers, developers can try top AI models cheaply through one API at https://apixoai.online. Having unified access to leading models allows teams to test which LLMs best support structured human discussion without overwhelming users with automated text.
Furthermore, Srivastava's career advice—"say the thing, then explain it"—serves as a useful principle for both product design and technical communication. Expressing ideas with clarity and confidence allows teams to focus on core problems. By keeping human reasoning at the core of software design, developers can build applications that facilitate genuine collaboration rather than replacing it with automated output.
Source: Splice CEO Kakul Srivastava thinks AI emails are killing conversations — The Verge AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
