The Nebula Awards spent one very chaotic day at the end of last year proving just how unresolved the question of AI still is in publishing. The Science Fiction and Fantasy Writers Association briefly opened the door to partial AI use in nominated works, with a disclosure requirement so voters could decide for themselves. Within hours, after backlash, they reversed course entirely and moved to a blanket ban: any work that used LLM tools at any point in its writing process would be disqualified outright.

I want to push back on that instinct, respectfully, because I think it gets the solution backward.

Trust the Reader, Not the Gatekeeper

SFWA's first instinct, before the reversal, was actually the right one: disclose the AI use, and let the people voting decide whether it matters to them. That's not a radical idea. It's the same principle that already governs almost everything else readers choose to buy. Nobody bans a book for using Grammarly, Scrivener, or a paid developmental editor. We trust readers to judge the finished work, not the tools that shaped it.

A blanket ban replaces that trust with a gatekeeper's judgment call, and it does it retroactively, applying to the entire writing process rather than the finished product a reader actually holds in their hands. If a story is good, it's good. If it isn't, no amount of disclosure or non-disclosure will save it in the marketplace. Let the sales numbers, the reviews, and the reader response be the verdict. That's a more honest form of accountability than a committee deciding in advance what's allowed to compete at all.

What Gets Lost in the Ban Conversation

The disclosure-versus-ban fight tends to focus entirely on award eligibility and prestige, which is a fairly narrow slice of the actual indie publishing world. What doesn't get talked about nearly enough is what AI has done for the authors who were never going to be in the awards conversation in the first place, because they couldn't afford to publish competitively at all.

Developmental editing alone can run well into four figures. Add a line edit, a professional cover designer, a marketing consultant, and audiobook narration, and you're looking at a genuine barrier to entry that has kept enormous numbers of talented, unknown authors out of the market entirely. Not because their stories weren't good enough. Because the on-ramp was too expensive.

AI tools have dramatically reduced that cost. Editing assistance, developmental feedback, cover concepts, marketing copy, and even voice narration are now accessible to an author working with a fraction of what those services used to cost individually. That's not a shortcut around quality; it's a bridge over the exact financial gatekeeping that traditional publishing has always relied on, and that self-publishing was supposed to dismantle in the first place.

For an unknown author with a real story and no capital behind them, that shift is enormous. It's the difference between a manuscript that never leaves a hard drive and a book that actually reaches readers who get to decide for themselves whether it was worth their time.

Disclosure, Not Prohibition

None of this means transparency doesn't matter. Readers deserve to know how a book was made if they care to know, just as some readers care about a translator's name or whether an audiobook was narrated by a human or synthesized. Disclosure respects the reader's ability to make an informed choice. A ban just makes that choice for them before the book is even written.

The market is already answering this question in real time, genre by genre, reader by reader. Let it keep answering. The books that connect will sell. The ones that don't won't, regardless of what tools sat on the desk while they were made.

Sources:
- Self-Publishing Advice Center (ALLi), "Podcast: Nebula Awards Reverse AI Policy"
- File 770, "SFWA Launches, Aborts LLM Tools Nebula Rule Change on Same Day"
- The Steampunk Explorer, "SFWA Retreats After Opening Nebula Awards to Works Created with Gen AI"