The fallout from Seattle Worldcon 2025 Chair Kathy Bond’s public statement attempting to defend the use of ChatGPT as part of the screening process for program participants now includes Yoon Ha Lee’s rejection of his status as a Lodestar Award finalist:
SEATTLE 2025 SOCIAL MEDIA. The original Seattle Worldcon 2025: “Statement From Worldcon Chair” post on Bluesky continues to be a magnet for criticisms of the committee, shaming, demands for resignations, a call for all panelists to reject their invitations, and ridicule of the Worldcon in general, which can be read at the link.
There are over 50 comments on the Seattle Worldcon 2025 Facebook post of the statement covering the same spectrum. These two are from writers who have commented here on occasion:
Back on Bluesky, Jasmine Gower asserts “Their own Privacy Policy does NOT give them permission to share your personal data (even ‘just’ your name) with genAI”, and says after contacting the con to object, they are getting a full refund of their membership.
OPEN LETTER. Jake Casella Brookins of Hugo-nominated Ancillary Review of Books posted a “Letter of Concern about LLM usage in Worldcon Panelist Vetting” which says in part:
… I also want to express that the disconnect between the concom and the larger SF community on this issue is, to me, even more concerning than the narrower technical decisions. The ethical, environmental, and practical issues with AI are loudly, widely, and routinely discussed in the science fiction community, with many artists directly impacted by AI plagiarism; community members of all backgrounds frequently voice their positions against it. Even very slight familiarity with this topic—on social media, as discussed at other conventions, at all levels of publication from professional journalism to personal blogs—would have warned against using AI for this purpose and predicted this community response. Whether or not Seattle’s vetting program was practically or ethically sound, the decision to use ChatGPT scripts, and the language in this disclosure, speak to either ignorance of or disregard for an intense opinion vocally held by a very large portion of the SF community.
I was—and still am—incredibly delighted and honored to be a Hugo finalist. And, as someone who rarely gets to meet other fans in person, and who can rarely afford to travel to Worldcon, I was incredibly excited to attend and participate this year. It’s personally crushing that, unless the concom takes major steps to address this controversy, this will be another Worldcon that will always have an asterisk next to it, another Worldcon that unnecessarily creates a lot of bad feelings and bad blood in the community. Whether or not it’s accurate to the situation, “the Worldcon where panelists were selected by the racist plagiarism machine” is going to be what it’s remembered for if significant steps are not taken, and quickly.
I urge the concom to take the reputational damage being incurred extremely seriously, and not to dismiss the practical concerns about how LLM usage affected panelist selection. I ask that you look at the response to this statement—on all channels—equally seriously, to see the level of anger, hurt, and division it is causing. In a year when the host country’s institutional bigotry is already significantly affecting who will or can attend Worldcon, it feels particularly important to set this right…
OTHER COMMENTS.
Alma Alexander wrote in a comment on File 770.
I am… currently mulling my options concerning the whole mess. I don’t particularly want to go the rest of my life and my career with “AI VETTED” hanging around my neck like a scarlet letter, especially after I’ve been so vocal in disavowing it, of distancing myself from it, in stating unequivocally that I do not want or accept the presence of AI anywhere near my creative endeavours. I am in contact with a number of other authors who feel the same way.
Shawn Marier, who runs Seattle’s film festival, stood up for the Worldcon’s use of LLM.
David Gerrold is another rare instance of someone who supports the committee.
More than one writer raised the spectre of ChatGPT’s reputation for racial bias in its results.
One commenter feels the skillset of Worldcon runners needs to be expanded.
The present controversy is also feeding on the discontent which follows when some applicants are not selected as panelists, and the various ideas that notability, awards, or longevity in the field should govern who is picked.
David Boop on Facebook.
… Let’s talk WC for just a moment. I haven’t attended many, usually because they happen close to Dragon Con, which is a place I do most of my business for the following year. But I wanted to do WC this year, not only because it’s a city I love, but I have a big release from Arc Manor Publishers coming out that, for the first time, might be something the Hugos might consider worthy of notice. This was important to me, being there, among my friends, my peers, my publisher. So, I submitted my application, expecting the best.
When I got my denial, it was from Sunny Jim, who I’d hoped had learned that I wasn’t that same naïve baby-author of the past. I also found it weird, having seen so many posts about international authors cancelling their memberships because their afraid to fly here (that’s a can of worms I don’t want to open here, and please, please don’t in the comments). that they didn’t have space for more people on programming.
But then it comes out that the programming staff didn’t even vet the authors until after their names through an LLM program, which is notoriously unreliable. I don’t know if I was kicked then, or when they did their “review” later, or if I made it to the final round before being voted off the island. There’s a lot that’s unclear about this process. All I know is that I’ve heard from friends that they got on programming with a lot fewer credits than I have, and don’t have a major, Hugo-worthy release coming out later this year. They didn’t attend NWC and did everything that was expected of them to earn their consideration. And now this frickin’ AI bullcrap after being one of the tens of thousands of authors to have the majority of their work stolen and loaded up into similar programs?
I know every WorldCon ConCom starts anew, and you can’t blame them for past mistakes. I agree. The problem is, the new committee is so focused on not repeating the mistakes of the last committee that they leave themselves open to new mistakes. This was a doozy!
I have a flight, I have a hotel, and I have a membership, but all the excitement is gone. There’s an empty feeling inside where once there were possibilities. Yes, I can still attend, network, promote. None of that has been taken away from me. But being on programming, showcasing everything you’ve learned and accomplished in the years since you last attended a WC is important. It’s validation. And yes, I still get that from Dragoncon, as well. And the many other cons I attend. And every time I sell something new, and every time I do the big shows like San Diego Comic Con, etc. But my heart wanted this. I debuted at Denvention in 2008. I got my current agent at Kansas City. I wanted this.
But it’s very possible a computer, not a living being, said No….
Discover more from File 770
Subscribe to get the latest posts sent to your email.
In related news, Metaphorically is being sued because of false assertions from its “AI”
Respect for Yoon Ha Lee. Withdrawing from consideration from a major award is a big step
David Gerrold’s comment feels like parody. Is he really taking the position that anyone who appreciates some uses of modern technology has to appreciate all uses?
I’m finding a certain amount of revelation about people who are defending this use of AI. I’ll be honest–I did not expect to be selected for virtual panels by this particular Worldcon for various reasons, including never being accepted by Norwescon despite applying for panels several times–even when I was on the board of a regional independent writers’ association.
Not surprised by some defenders, but others? Definitely eye-opening, and I’m keeping track.
I finally broke down and asked ChatGPT “who is Joyce Reynolds-Ward”? The answer was word-for-word from my author bio. Nothing more, so at least whatever vetting happened didn’t appear to impact that (and at least one source was from one of my books that they stole!).
That’s my one and probably only usage of the damned technology. I’m also not thrilled by defenders casting it as a “Luddite” issue and talking about its use as a valid tool. It was still developed using theft of materials, and is an intensive use of water and energy.
Mind you, I see valid usages for this tool, primarily with closed databases that are huge (think inputs from agricultural field mapping for one, which can be very useful combined with smart pivot sprinklers and smart tractors/drones for appropriate application of water and other inputs, down to square centimeters or so). It has the potential to revolutionize precision agriculture technology.
But that’s a far cry from stealing the work of creatives, much less using it to vet convention panelists when I’ve already seen multiple examples of biography errors being posted on social media. That’s the issue that the defenders appear to be dodging.
And, interestingly, I can’t see the particular Facebook post referenced here. I see other things but not this one.
Having to use them because of my field of paid work there’s a lot I would say on this but using them to answer narrow questions like “from the public sources would person X be likely to cause us a problem for our CoC (link)” yields decent results and a list of potential issues that, as I understand it, led to a manual check.
Also, reading what the con has said, this was a vetting check AFTER they had provisionally identified people for the program wasn’t it?
“All Fandom Plunged Into War” for the 21st Century…
I can apply one of my clichés: “Here we are in the future, and it sucks…”
David Gerrold’s comments are a disappointment. What data is being gathered and sorted? How? What biases are being introduced and how are they being watched out for? Blandly generic words do not illuminate.
Using an LLM, something that is still a raw unhealed wound in the community, seems a poor decision. Techbros are still gloating about generative AI replacing the world’s creatives and putting them out of work. Why invite trouble by using a mediocre sorting mechanism with a putrid reputation in the community it’s being applied to?
Two and a half years ago I moderated a panel on the then-breaking news about AI-generated imagery and LLMs. At the time there was a lot of hype and a lot of anxiety. I was hoping the panel could cut through the hype and if not allay, at least focus the anxiety.
I delved deep into the math and research before the panel, to make sure I understood as much about what was going on as I could. I was impressed by what I saw, but not in the way the people selling AI are. I thought generative AI was very clever and very interesting, but its underlying math means it is in no way creative and can never be so. And the language models have no relation to fact.
And no, it is never going to be intelligent. That’s pure fantasy.
Not all the panelists understood the mechanisms behind generative AI, and the most enthusiastic about it knew the least.
Any interest I had in the mathematical cleverness went out the window when the scale of the theft of the intellectual property became apparent. Generative AI is hopelessly tainted by thievery, and no amount of business-suited respectable indignation can change that it’s all built on a massive pile of theft, forgery and plagiarism.
Under the circumstances, given the trauma of having their works stolen to feed a monster explicitly and openly meant to undercut and replace them, using that monster to sort which of the sci fi community are worthy of being on panels seems an unforced error, an absolutely unnecessary cruelty.
@rcade: Yeah, I was waiting for the punchline.
“Ohell, I find it ironic for anyone to be objecting to the use of modern technology — you know, the technology that gave us the hydrogen bomb and drone warfare, the technology that put the hole in the ozone layer, the technology that lets corporations surveil you 24/7, the technology that is built on theft of the work of human beings on an unimaginable scale, all that.”
@rcade: Yeah, I was waiting for the punchline.
“Ohell, I find it ironic for anyone to be objecting to the use of modern technology — you know, the technology that gave us the hydrogen bomb and drone warfare, the technology that put the hole in the ozone layer, the technology that lets corporations surveil you 24/7, the technology that is built on theft of the work of human beings on an unimaginable scale, all that.”
@OGH: No idea how the double post showed up.
Why on EARTH would they need AI for something fandoms have been doing for DECADES?? Either AI brainrot is actually making people stupider and lazier or these people are paid by AI companies. In any case, they should be kicked out of the convention and everything re organized by HUMANS.
ima: I don’t think they need AI, however, you are vastly exaggerating when you treat the task as being the same as it was decades ago. Worldcons have only had Codes of Conduct for maybe a decade, and for even less time than that have any committees tried to screen out some of those whose conduct or social media posts conflict with their CoCs. And now that virtual programming opens the gates to people who won’t attend in person they get a huge number of requests that have to be considered — 1300 this year. When I organized the Worldcon program — several decades ago, coincidentally — I probably had to deal with less than 200 participants.
The ethical and actual usefulness issues aside, surely the optics alone should have made them rethink this plan? I remember a prof in college teaching about decision making in business and how there’s a number of different factors that go into a decision and one of them is optics/how it will be received and that sometimes even if the other factors are telling you to go for it, the optics are bad enough that you need to do something else. This doesn’t seem to be a concept the con organizers are familiar with.
I have yet to see a Worldcon that has not been attacked by fans for not living up to their own expectations. I think it must be part of the process. But if anything, it is evidence of how much we care about the integrity of our traditions.
Over here, I am operating under the naive assumption that any Worldcon committee has the best interests of the fannish community at heart. (This has not always been the case, but I like to believe it because I’m old-fashioned that way.)
I was not in the room for any of this and therefore cannot comment knowledgeably on the specific decisions or methodologies of the Seattle Worldcon committee. (And I regret being dragged into this by File770. I had my fill in 2015 and have tried to remain apart from fannish politics ever since.)
What I said above and what has been quoted here (without my permission) was not an endorsement of support for anyone or anything, merely an observation that none of us know enough about this to comment knowledgeably. Myself included. I remain an agnostic, willing to learn more.
But again, I do assume that the Seattle Worldcon committee did not act out of malice or even thoughtlessness. I assume that they thought that this would be valid way to deal with the challenge of sorting through1300 possible panelists — let the software do its thing, but the final decisions will still be made by humans. That’s how I read it. (That’s how I sort my to-do list. The software does its thing, then I decide what I want to do. So maybe I’m biased that way.)
Was it a mistake to use Chatgpt? I don’t know. I do know that Chatgpt is a lot more accurate than it was a year ago. But is it accurate enough? There is plenty of evidence to suggest that it is not.
Since writing the comments posted above (posted here without my knowledge or permission or any sense of the larger context in which they were written) I have now seen several very thoughtful analyses from people who do understand how Chatgpt hallucinates. I do think those informed opinions are worth paying attention to. (I have my own experience with computers — a half-century of experience, which is why I check everything.)
But please do not assume I am either supporting or condemning anyone or anything. I am not. I would like to believe that the Seattle Worldcon committee has been acting in good faith.
Yes, I would also like to know more specifics about the exact methodology. Are people overreacting? Or are the concerns valid? I do respect the opinions of those who know what they are talking about. That’s how I learn better.
Like it or not, we are entering a new era of human interaction, where we are offloading too much of our personal responsibilities to software that has too often been improperly vetted. I have more than once run into, “But the computer says…” or worse, “I’m sorry, there isn’t any way for me to fix this, it’s all in the computer…”
Any system that does not allow for direct human supervision, control, and if necessary, intervention is inherently biased and flawed. (I think I wrote some books about that too.)
I haven’t been on a con-committee in a long time, but my experience then (and remains so today) is that the convention is for fun. Otherwise, why do it? Every con-committee tries to use technology to simplify process. Some succeed, some do not.
Let me add this — the suggestion has been made that it might be time for a permanent organization to manage some of the mechanics of the Worldcon. I’m not sure how that would work, but I know the intention is to establish some consistency and help various con-committees be more efficient. But that’s another discussion.
David Gerrold: Your post is marked public. I don’t need your permission to quote a public statement of yours. I also reject your rhetorical claim there is some greater context in which this post appeared that would alter the reader’s interpretation of it.
@David Gerrold
A public post is fair to quote.
Perhaps you didn’t – and maybe still don’t – realise how flippant, misleading and dismissive your original post comes across. Oh well if people like vaccines then it’s weird to object to a completely different technology that was built by stealing their work, regularly hallucinates, and reinforces and regurgitates numerous bigotries – come on, now.
Taking a little more time to understand the objections, which were not, as you implied, thinking Worldcon was cheating somehow by using software to speed things up, might have helped.
However, if you think there is some missing context that improves how your facebook post and comment would have come across, you may feel free to indicate what it was.
I am frankly a little creeped out by some of these reactions. I am appalled by the use of AI but I also wonder if populating panels with people who will proceed to place their book in front of them at the panel leads to an experience I want to share. People seem to think they are owed a place on a panel because they have “credits”? Because they have a book coming out?
I have been to cons where panels were almost exclusively “let us tell you about our new books”, and I have been to cons where the majority of panels were people who had never had a word professionally published talked about depictions of green slime in fanfic, and while great fun can be had at the bar in both, it is the panels in the second one I enjoy.
I understand the anger of using ChatGPT, but I am a bit confused about what was done with it.
I understand it that way that it was done as a googlesubstitute to help the con identify, who are those people and are any of them problematic (which has quite a few problems, granted)
Some of the people quoted and in the discusion seemed to belive that ChatGPT did decied who got invited and put on what panel (which I can’t believe any sane person could exspect to have any good results except per exident).
I am not wanting to defend the use of ChatGPT, but I do hope that most of the people stating that they were rejected, where rejected by human beeings for the old problem that there are to many potential penalists.