We can train an AI system to behave in ways we prefer. The harder question is how we would know the training worked, and whether it will hold when things change.
Hi, I'm Arsh Rai.
I write documentation for complex enterprise software, and design the systems around it.
About
about me
For over a decade, I’ve written and led documentation for enterprise SaaS products across identity management, cloud infrastructure, and IoT, including API and CLI documentation. Today, as a Senior Technical Writer at Okta, I work on documentation architecture, publishing workflows, and AI-enabled systems that make complex information easier to create, maintain, and use.
Alongside hands-on writing, I increasingly design the systems around it: leading documentation for product launches across multiple partner teams, redesigning the information architecture, and building AI workflows where a writer approves every change.
When I’m not writing, you’ll find me recording music, exploring the city on my bike, or wandering the woods.
Recent Posts
Human in the Loop, but Is Anyone Looking?
Human review is often treated as the fail-safe for AI agents. But when a reviewer can't understand what happened, or doesn't have the time to check, approval becomes a formality—and the controls that don't depend on catching the mistake start to matter more.
Who Is Responsible for Keeping AI Systems Safe?
When an AI system causes harm, responsibility rarely sits in one place. A look at the operational and security side of AI safety—who controls which safeguards, and who is in a position to reduce the risk.
