No, I do not publish AI-generated content.

Everything you read, see or hear on Broad History is human-made. I write every line and speak every word. The art isn't great because I can't make it or (yet) pay someone to make it. Until then, I use public domain images and royalty-free music. (Become a member and help me hire folks.)

Why not?

First, it's ugly. The illustrations are bland, the writing is stale. Yes, it's gotten better but it is designed to revert everything to the mean. AI can be a helpful tool, but anyone who straight publishes what it regurgitates is making an inferior product.

I think when it comes to art and storytelling, being human is our unfair advantage. We don't write to have something written; we write because we have something to say. We don't read text; we read inside another human's mind and we hope it somehow feels a little bit like what's in ours. The second I realise something is synthetic, even if it is technically "good", I lose all interest. What's the point? Maybe that will change but for now I don't think that engineering a clever prompt counts as a creative act. Or if it is, the creation is the prompt – not the AI's interpretation of it.

I wouldn't mind it so much if gen AI was just another tool. But the economics of it are profoundly unfair and destructive. Many gen AI models are built on the theft of creatives' work and I will not condone it, let alone pay for the privilege. If we creatives don't stand up for one another, who will? I think of it the way I think of GMOs; it's not the principle of it that bothers me so much as the reality of how it exists in the economy, who profits and who gets screwed. We are quickly destroying the already precarious economics of creative work, and we will all lose out. Starting with the AI barons who'll run out of good stuff to steal.

Finally, it's important that you trust me and my work. I'm trying to tell you true stories about the past. Why muddy the waters with images that may or may not be real? You're grown-ups, I don't need to show you an AI Joan of Arc burning at the stake to get or keep your attention. Those "historical" AI videos you see on TikTok are so full of errors and mistaken assumptions about what the past looked like, they're not broadening access to historical knowledge, as their proponents pretend. They're a giant misinformation machine.

Where do you use AI?

I use it where it's good and reliable. Ever since making my first podcast in 2020, I've used AI transcription and text-based editing to make my episodes, first with Descript, now with Riverside. The software also places chapter markers, cuts out ums and ahs, identifies good excerpts for social media and autogenerates a first draft of transcripts and show notes. It can improve sound quality and cut out background noise. It does not create sound or image out of thin air, and none of it gets to you without going through me. It just makes my work as a one-woman-team much faster. I use Granola to transcribe and summarise all my meetings. I have some automated workflows and I've also used Claude or Cursor to help me code parts of the site that were resisting me. Next I'm trying to build something that will crawl all nonfiction publishing catalogues, alert me of new relevant women's history books and draft my outreach to publishers and authors. This kind of work takes me hours to do manually every publishing season. See, it's not all bad. As long as humans stay in the driver's seat.