Thoughts on the Nebula Awards, LLM usage, and where we draw the line

Yesterday the Science Fiction and Fantasy Writers Association updated the rules for the Nebula Awards, one of the SF/F genre’s premier honors. The first rule change gave Nebula eligibility to works partially created by large language models (LLMs), the tools most people know as generative AI. But after an immediate outcry by members, SFWA quickly rescinded that change and updated the Nebula rules to prohibit all LLM usage in the “writing process.”

You can read a deep dive into all this in my original post from yesterday.

I appreciate SFWA listening to their membership and moving so fast on this issue. I also like that the newly updated Nebula Award rules focus on honoring human creativity and don’t allow LLM usage during the writing process.

However, there are already complaints that the second rule change went too far and could trip up authors who use search engines powered by LLMs for research, or use spelling and grammar checks that are likewise backed by LLMs. I believe these issues are what SFWA was trying to address with their original rule change, even if the wording they used would have caused far more damage than intended.

LLMs are a very divisive issue in our world, not only because the billion-dollar companies creating these programs trained the tools on works stolen from writers and artists. My own fiction was stolen to train generative AI programs. I refuse to use these tools in my writing, not only because of this theft but also because the tools are not actually creative and defeat the entire point of storytelling.

That said, it’s important to remember that these generative AI products are being forced down everyone’s throats by major corporations. If you use any online search engines or computer products these days, it’s likely you’re using something powered by or connected with an LLM. Because of that, we must be careful that writers who use word processing and research tools with LLM components aren’t unfairly disqualified from awards like the Nebulas or attacked by readers and other writers. We also have to be careful about not turning this into a witch hunt against writers, dissecting their every story to see if there is any LLM usage no matter how minor.

Here’s how it works for me: if a writer uses tools like spelling and grammar checks, speech to text or other accessibility tools, translation programs to read sources that aren’t in their own language, or search engines for research, we shouldn’t hold any of that against them even if an LLM powers those tools. The same if they use a writing tool like Google Docs that has an integrated LLM. As long as the writer isn’t using the “Help me write” prompt in Google Docs then Google Docs can be seen as just another word processing program.

My reasoning for this is it’s difficult for writers to parse out everything that’s behind today’s standard writing tools. Plus these tools are ones writers have been using for several decades. It’s only relatively recently that companies began adding LLMs to these tools. Expecting writers to suddenly return to writing with paper and pen because LLMs have been added to tools already widely used in our world is unrealistic and unfair.

But if a writer uses LLMs to craft or write their story, even partially, or copies text from LLM tools directly into their stories, then that shouldn’t be allowed in the Nebula Awards.

Again, I appreciate SFWA for listening to their membership on this rule change. To SFWA’s credit they also appear to be listening to the concerns people are raising about LLMs connected with standard writing tools like spell check and search engines. My hope is that SFWA will clarify the Nebula rules in this regard in the coming weeks.

The rules behind the Nebula Awards are both legalistic and inspirational. The legalistic aspect is used for running the awards every year. However, the inspirational aspect can be even more important because it’s what SFWA officially says through the Nebula Awards to writers and readers across the world. This inspirational aspect is SFWA declaring, “We are honoring human creativity, not works crafted wholly or in part using LLMs trained on stories stolen from other writers.”