Genre Grapevine on the 2025 Seattle Worldcon AI Fallout
If you like my genre reporting, please consider doing a paid subscription or backing my Patreon.
Last month I received exciting news: the report Chris M. Barkley and I wrote about censorship and exclusion in the 2023 Hugo Awards was a finalist for this year's Hugo Award for Best Related Work.
Despite the irony of such a report being a Hugo finalist, I was thrilled by the honor because it recognized the impact our reporting has had on both the overall SF/F genre and Worldcon. Since the report came out, numerous people in the genre have proposed ways to improve Worldcon and the Hugo Awards, with some of these proposals moving forward at last year’s World Science Fiction Society business meeting. And last year’s Hugo Awards also ended up being among the most transparent in genre history.
Ever the eternal optimist, I hoped Worldcon was finally turning the corner from the regular controversies that beset both the convention and the Hugo Awards.
Turns out the answer to that is a big "Nope!"
In recent weeks, I heard unconfirmed rumors that Seattle Worldcon was using generative AI systems – more correctly known as large language models (LLMs) – to vet people for consideration in Worldcon programming. Again, being the eternal optimist I hoped this wasn't true, especially since such an LLM would have been illegally trained on the works of so many of the very people who'd be attending Worldcon.
Nope x 2.
On April 30, Seattle Worldcon Chair Kathy Bond released a statement confirming an LLM had indeed been used to vet the more than 1,300 people who’d applied to be program participants.
"In the interest of transparency, we will explain the process of how we are using a Large Language Model (LLM). We understand that members of our community have very reasonable concerns and strong opinions about using LLMs. Please be assured that no data other than a proposed panelist’s name has been put into the LLM script that was used. Let’s repeat that point: no data other than a proposed panelist’s name has been put into the LLM script. The sole purpose of using the LLM was to streamline the online search process used for program participant vetting, and rather than being accepted uncritically, the outputs were carefully analyzed by multiple members of our team for accuracy.
"... In order to enhance our process for vetting, volunteer staff also chose to test a process utilizing a script that used ChatGPT. The sole purpose of using this LLM was to automate and aggregate the usual online searches for participant vetting, which can take up to 10–30 minutes per applicant as you enter a person’s name, plus the search terms one by one. Using this script drastically shortened the search process by finding and aggregating sources to review.
"... Using this process saved literally hundreds of hours of volunteer staff time, and we believe it resulted in more accurate vetting after the step of checking any purported negative results. We have also not utilized an LLM in any other aspect of our program or convention."
As someone who has heavily covered generative AI for several years along with the impact these systems have on artists and writers, I wasn’t convinced using an LLM actually saved Seattle Worldcon any significant volunteer time. After all, the NY Times recently reported that the error rates of "newer A.I. systems were as high as 79 percent." This means that nearly every bit of info returned by the LLM would have to be verified, negating any time savings.
Unless, of course, all that information wasn’t verified.
In addition, using LLMs to vet panelists is a powerful slap in the face of the very artists and authors who attend Worldcon and have had their works pirated to train these generative AI systems. My own stories were pirated to train LLMs. The fact that an LLM was used to vet me really pisses me off. And you can see similar anger from many other genre people in the responses to Kathy Bond’s post, with more than 100 comments ranging from shock at what happened to panelists saying they didn’t give Worldcon permission to vet them like this.
The anger from the Worldcon and genre community was so great, Bond released an apology two days later, saying the "initial statement on the use of AI tools in program vetting was incomplete, flawed, and missed the most crucial points. I acknowledge my mistake and am truly sorry for the harm it caused."
In the apology, Bond also pledged to release a statement by today with a "transparent explanation of the process that was used, answers more of the questions and concerns we have received, and openly outlines our next steps."
Sounds good, right? Everything now squared away?
Nope.
Yesterday, the administrator of the 2025 Hugo Awards released this message:
"Effective immediately, Cassidy (WSFS DH), Nicholas Whyte (Hugo Administrator) and Esther MacCallum-Stewart (Deputy Hugo Administrator) resign from their respective roles from the Seattle 2025 Worldcon. We do not see a path forward that enables us to make further contributions at this stage.
"We want to reaffirm that no LLMs or generative AI have been used in the Hugo Awards process at any stage. Our nomination software NomNom is well-documented on GitHub for anyone to be able to review. We firmly believe in transparency for the awards process and for the Finalists who have been nominated. We believe that the Hugo Awards exist to celebrate our community which is filled with artists, authors, and fans who adore the works of our creative SFF community. Our belief in the mission of the Hugo Awards, and Worldcon in general has guided our actions in the administration of these awards, and now guides our actions in leaving the Seattle Worldcon."
The statement was signed by Cassidy, Nicholas Whyte, and Esther MacCallum-Stewart. It’s worth noting that Nicholas Whyte stepped up as the overall 2024 Hugo Awards administrator when the previous administrator resigned due to involvement in the censorship of the 2023 Hugos. Cassidy also helped run last year’s awards, which were extremely transparent and well-managed.
I want to emphasize how respected and trusted Nicholas Whyte is in the genre community. Think of Whyte as similar to Cobra Bubbles in the beloved Disney film Lilo and Stitch, where he famously says, "I am the one they call when things go wrong, and things have indeed gone wrong."
When the one you call when things go wrong says “We do not see a path forward,” that tells me some serious shit went down and there is resistance to fixing said shit among the Seattle Worldcon staff.
So what does all this mean?
First off, Kathy Bond confirmed to me last night that Worldcon will be releasing a statement today about all this. I sincerely hope that Bond and the Seattle Worldcon staff realize the precarious situation Worldcon has found itself in and will take steps to not only illuminate what happened but fix things going forward.
Second, the people running Worldcon need to understand how pissed off people are about all this. Elizabeth Bear has now withdrawn from Worldcon over "the ethics of their use of genAI to vet panelists." Fran Wilde has done the same. Yoon Ha Lee also withdrew the novel Moonstorm from consideration for the Lodestar Award, where it was one of this year's finalists.
Such withdrawals could easily mushroom if Worldcon doesn’t decisively fix this problem in the very near future.
And all this is taking place as many people from around the world are already hesitant about attending a Worldcon in the United States, where travelers and immigrants are continually subjected to arbitrary and harsh treatment including imprisonment under Trump’s new immigration rules. I’ve seen a bunch of comments from genre friends outside the USA saying they will not attend this year's Worldcon because they feel unsafe visiting the country. Echoing these reasons, long-time genre fan and author Gary Westfahl recently wrote an essay in File770 calling for the 2025 Seattle Worldcon and the 2026 Los Angeles Worldcon to be cancelled or moved.
It's not hyperbole to say that if Seattle Worldcon doesn’t get their shit together, this year's convention could totally fall apart, which would badly impact future Worldcons.
That said, it's not all doom and gloom. The finalists for the Hugo Awards have already been announced, with voting ongoing and running through July 23rd. If the protest resignations of Whyte and the other Hugo staff had happened earlier, the awards would have been seriously endangered (which is likely why their resignations only took place now). So even if Whyte and the rest of the Hugo staff don’t step back into their roles, there is a reasonable expectation that the award process can be completed.
In addition, the Seattle Worldcon has taken some positive steps in recent months, which shows that they can be responsive to the needs of the genre community. For example, for the first time ever, this year's Worldcon will host its business meeting virtually prior to the convention. This will make the business meeting much more open and democratic than it has ever been. Since the business meeting is where the rules and bylaws that govern Worldcon are debated and voted on, this change potentially allows for new groups of genre fans to influence how Worldcon will function in the years to come.
There is a very short window for Seattle Worldcon to fix their current problems. The eternal optimist in me thinks they can do this.
But we'll see. I hope I won’t be writing another "nope" in a few weeks.