This is my AI policy as of August 2026. I’ll revise it as the landscape changes.

I Do Not Generate Art With AI

Every word, image, video, etc that I publish is written/painted/created by a human, and not an AI model. This means everything—book content, blog posts, social media posts, Discord messages. The only possible exception would be when I’m writing about AI and including a sample of what it creates.

This is both a moral position about the role of AI in society and a decision about how I can meet my own goals of social connection and artistic growth.

It’s extraordinarily shitty for AI models to hoover up every artist’s work without permission and then use it to build a put-artists-out-of-work machine. I am in support of legislation that would require the creators of AI models to get permission before including copyrighted works in their training data.

I also want to connect with readers through my writing. That only works if the writing I put out genuinely represents me. If I were to put AI writing or other art out there, I’d be shortchanging myself as well as the reader.

Lastly, I’m trying to grow here. I am learning how to excite, tantalize, arouse, and horrify my readers. It’s hard, and that’s what makes it rewarding. Letting an AI write for me would be as pointless as letting a robot go to the gym for me.

Exercise Extreme Caution With AI-Assisted Research

When I have a research question, step 1 is to Google it. I’ve been doing that for decades. And these days the first thing you see after running a Google search is a summary that’s just as AI-generated as Claude or ChatGPT (which I’ve also tried and abandoned). Usually the general direction of its answer is reasonable, but then about 25% of the details underneath are misleading in some way.

So you have to chase down those details to a trustworthy source—a human with the motive and ability to get the answer right. For Sumerian translations, for example, that means the Electronic Pennsylvania Sumerian Dictionary. For questions on Victorian customs I’ve found historical fiction writers like Gail Carriger and Mimi Matthews to have incredibly helpful blogs, which then link back to primary sources in places like the Internet Archive, Project Gutenberg, etc.

Since I’m writing fiction, I reserve the right to tweak historical events and places to work with my story—like my creation of a fictional Chebsey Manor—but those should always be intentional choices, not AI hallucinations.

Be Gentle With People

I don’t want to read a book that someone generated with AI. But even more, I don’t want to destroy someone online with false accusations of the same. Cancel culture is purity culture, and a tool of the Jealous God. I don’t worship there anymore.

I may refer to the controversy of the day as a jumping off point for exploring new questions or nuances about how people are using AI, but this is not the place for witch hunts.