AI Personal Policy 2025
I’ve been using Large Language Model (LLM) tools since 2023. The technology is fascinating, and using LLMs can be an amazing amount of fun (and occasionally extremely frustrating). I’ve taught myself a bit of gravitation by walking through Scharzchild’s solution to the Einstein field equations, written up an entire suite of organizational documents for a startup, and produced a fair bit of software.
What I do not use AI for is writing. My writing is mine.
My personal policy as of this moment, and through 2025:
-
Unrestricted Use: programming and documentation around programming artifacts such as command line clients, web applications, and similar. It makes sense to leverage the ecosystem in this context, I’m hardly a special snowflake. LLMs are also remarkably fun for ideation and creative inspiration. This doesn’t feel much different than using books and articles for inspiration, and has the benefit of interactivity.
-
Restricted Use: Professional documentation around policy and process in the workplace. AI/LLM tools are excellent for generating generic boilerplate which fit multitudes of scenarios, but the resulting artifacts invariably need scrutiny to ensure 1. they are at the least not factually incorrect, and 2. adding, deleting, and rewording content to fit the specific context.
-
Not Used: While I may not be a snowflake, I do have a voice which is reflected in my writing. Writing is a difficult skill to learn, and to me worth continuing to improve at. I may use AI/LLM tools as editorial assistants for spelling and grammar, and possibly flagging badly worded passages. It is not difficult to prompt a chat client to refrain from rewording.
These are for 2025, and this article may be updated and expanded throughout the year. Who knows what 2026 will bring? The rate of technology change is accelerating!