Responsible
AI Principles
These principles guide how we build AI at Splice.
Technology should amplify, not replace, human creativity.
A great tool gets out of the way and lets you make something only you could make. That's the standard we hold AI to. We build sound-first, which means the starting point is always a human-made sound, and the creative decisions stay with you. AI should give you more ways to express what's in your head, not make the decisions for you.
Human-made art deserves recognition, protection, and respect.
Every sound in Splice came from a real person. A musician, a producer, someone who sat down and made something. That human origin matters to us, and it shapes how we build. We only train on sounds we have the rights to use. When those sounds contribute to AI-powered features, the people behind them get paid. And we won't ship anything that chips away at the value of what human creativity is worth.
Trust, collaboration, and transparency are essential in using AI.
We talk to creators before we ship, not after. Soundcheck is how we do that at scale: real musicians and producers inside the process early, stress-testing ideas, shaping features, telling us when something isn't working. We're also committed to being straight with you about where and how AI shows up in Splice. No buried footnotes. No vague reassurances.
Human oversight is necessary to protect culture, identity, and community.
Generative AI can cause real harm if nobody's watching. We review what goes in and what comes out. We maintain guardrails against misuse and cultural misrepresentation. The sounds that come through Splice carry culture, history, and identity, and we take that seriously.
We believe in innovation guided by long-term responsibility.
We're not chasing AI for its own sake. We're trying to make Splice better for the people who use it every day: better sound discovery, a deeper and more diverse catalog, a smoother path from idea to finished track. We're also working with our partners to keep an eye on the environmental cost of what we run, and to reduce it where we can.