On hearing the truth spoken to power.
I attended a conference earlier this year. It’s as much a small community as a conference, in many ways. I like to think that I’ve made some positive contributions to the conference, and to the community, over the past years.
Something happened this year; an interaction that I perceived as misogynistic and profoundly unjust. I’ve debated whether to write about this for a while. Finally, I realised that I was being held back by fear. If this had happened in any other context, I wouldn’t feel this fear. But the community in question is special to me and I’m afraid to find out that I don’t really belong there.
But if I can only belong by keeping quiet about what I perceive as injustice, then I don’t belong there in any case.
You are, quite obviously, getting only my side of this story. Other people involved would have a different version, and some of those versions would likely be much less flattering to me. But I hope that those versions would not gainsay the facts I’m going to outline, even if they would disagree about how those facts should be interpreted.
On its surface, the story is about my opposition to the “AI” industry—specifically to the enormous commercial LLMs such as ChatGPT.I’m going to use the term “LLM” as a shorthand for these commercial models. I understand that the underlying technology isn’t synonymous with these models, and that individuals can train their own LLMs or run local versions of commercial LLMs. But I’m only talking about the large commercial models here. I feel strongly that these LLMs have unethical choices built in to them. They cannot be used ethically because they embody ethical violations. And when you use an LLM, you are complicit in the creation of its next iteration, because you are providing it with new training data.
To paraphrase what a friend said to me, these LLMs “are like a hammer whose metal parts were strip-mined using slave labour, leaving behind toxic sludge in the groundwater, and whose wood comes from endangered tree species grown on land stolen from indigenous people”. Even if you need a hammer (and often you don’t really need a hammer, though the hardware store owner insists you do), is that a hammer you’re willing to use?
But this story isn’t really about LLMs. It’s about technology communities, and how we can make them welcoming, equitable, ethical spaces. It’s about how we react when somebody speaks their truth to our power.
As we know, the major tech companies are notorious for making unethical choices regarding user privacy, worker rights, environmental protection, mitigating bigotry and bias, and a host of other issues. As smaller tech communities, I believe we have a duty to do better, precisely because our larger counterparts are doing so badly. We have a duty to offer an alternative model of technology which centres morality and humanity.
I want to talk about how we centre morality in our tech communities. But first, I need to tell you what happened at the conference. This is quite long; I’m trying to be as open as I can about how I behaved, and why, and why I found the response I received so troubling.
The conference, which was fully virtual, featured a talk about learning to love using LLMs in your professional workflow; the published abstract for the talk promised a discussion about the ethics of LLMs following the talk. Some of these ethical issues were featured towards the end of the speaker's presentation, though his approach to them didn’t seem excessively thorough.
So, after listening to a few other attendees ask their questions in the Q&A following the talk, I asked something along the lines of: “You mention a few ethical concerns, but don’t cover many others. Was that choice deliberate, or do you not know about them?” The speaker said some things I don’t now remember in detail. I do remember asking “What sources did you use in your research into the ethical issues with AI?”
The speaker answered “I asked the AI”.
I laughed, expecting a bit more follow-up to the joke. I was somewhat irked when there wasn’t any. I remember saying, sarcastically, “Did you think to ask any other source than the AI about the ethical problems with the AI, or do you just not really care?”
I’m told that my “do you just not care?” question raised some eyebrows, and came across as sharp and rude. I suspect that’s a fair interpretation. I’m autistic, so recognizing how my tone comes across to others isn’t my strong suit, though I try to be aware of it. A couple of other attendees later messaged me to thank me for pressing the speaker about this issue, and didn’t mention my tone at all, which makes me think that I wasn’t so rude as to be universally shocking. But I was, undeniably, feeling provoked by what I thought was a disingenuous answer to a serious question about a topic that the speaker had claimed a willingness to discuss.
The Q&A session moved on. I’d never met the speaker before, but he struck me as friendly and open, a confident presenter with decades of tech experience and a lot to share. At the end, a bunch of us stayed on the videocall, and the speaker and I had a good chat about cats.
The next day, I received an email containing an official reprimand and a final warning from the conference chair, threatening to throw me out if I displayed “another episode of disrespect toward a conference participant”. The reprimand included an excerpt from a letter of complaint the chair had received from an attendee, with the chair’s own comments in addition. I want to note that I have no reason to believe that the speaker himself was the complainant, particularly since I had a pleasant email exchange with him after the session had concluded.
Specifically, I was accused of harassment, of “spraying vitriol”, and of acting against the “ethos” of the conference. I was also accused of “sidetracking” the conversation to pursue a personal “crusade” on a topic “not directly relevant” to the discussion. I was called “hectoring”, “aggressive and confrontational” (by the attendee) and “forceful and disrespectful” (by the chair). I was not offered the opportunity to respond or to put my side of the case.
The accusations were based largely on the interaction I’ve just described, but also mentioned two comments I made in the conference’s text chat channel. The text channel runs concurrently with the video channel, and attendees will comment on points of interest in the session, share links and references, and chat. It wouldn’t work for all communities, but I’ve noticed that techies often like to multitask, and having a chat parallel to the main session adds a lot to the conference.
My first comment, during a vendor presentation, said that using LLMs to generate alt-text is a “massive accessibility fail”, and expressed distaste that a vendor would propose using LLMs for translation. I stand firmly by these comments.
The second comment was during the talk I’ve just described, when a participant told me that ethics wasn’t a relevant topic, and I responded that if they meant “shut up so that we can have a nice tech talk and get excited”, they should complain to the organizers for allowing my question. I was bemused by this resistance to my question about ethics, and annoyed to be accused, incorrectly, of derailing the conversation by raising an irrelevant topic.
There was general discussion of the ethical problems with LLMs in the text chat for a number of sessions, with the vast majority of participants expressing concerns and criticisms. I was not a lone voice banging on about this issue; it evidently engaged the interest of a large number of participants with concerns about LLMs, including, notably, two of the conference co-chairs.
Although the conference chair’s endorsment of the complaint that I had “sidetracked” the discussion confused me, I was more troubled by the way my behaviour had been characterized by both the complainant and the chair. To describe me as “hectoring”, “aggressive”, and “forceful” (apparently this last was not intended as a compliment) displays classic misogynist tropes about women’s speech See e.g. the books by Cameron and Manne in the Further Reading section below.. This was, let’s remember, one fairly short and contained interaction with a speaker. I was sarcastic, yes, and frustrated because the speaker’s flippancy struck me as somewhat disrespectful—ironically enough, given that I was now accused of disrespect. But that was the extent of it.
As a speaker in previous years, I have been talked over, shouted down on one occasion, shushed, had ideas treated with sarcasm or skepticism, and had the Q&A of my own talk derailed by two attendees who went off on a tangent about something of interest to them. This doesn’t define my experience of the conference, which I have generally found stimulating and engaging. But I have certainly not always found the tone of conversation to be politely respectful, and Q&A sessions often wander into unplanned areas of discussion. In particular, I have (fairly rarely, thankfully) had interactions which displayed unthinking misogyny; I don’t think I would have been shushed if I were an older white man.
And yes, almost all of the behaviours I describe—the interruptions, the raised voices, the hijacked conversations—involved older white men.
So why did my behaviour enrage the conference chair so badly that they accused me of harassment, vitriol, and aggression? And why didn’t the behaviour of those older white guys towards me enrage the chair? Why does a single interaction characterized by sarcastic rudeness mark me as potentially unfit to remain in the community, when there are male attendees who are habitually abrupt, impatient, or dismissive in their interactions?
Why was my “forceful” speech about probably the most crucial ethical problem for technologists today such an obscenity?
I think I have an answer, even if it isn’t the whole answer. I think this stems from deeply ingrained beliefs about what’s acceptable from women in public discourse, combined with a commitment to civility as a trope in white American culture. Much has been written about both topics, and I won’t attempt to summarise the literature hereThere are references at the end of this post, if you’re interested in reading more.. I think the conference chair overreacted so badly to a woman’s vehement criticism of an older white man because cultural norms magnify that transgression, while minimizing the same kind of speech from men towards a younger woman.
A final wrinkle comes from the specific, formal accusation of harassment made against me by the complainant, and upheld by the chair. I was mystified that one sharp interaction with a speaker, however rude I might have been in the moment, was characterized in this way. The use of the word “harassment” is designed to make sure that the accused knows just how seriously they’ve trespassed against social, and possibly legal, norms. It tells them that they have transgressed, perhaps unforgivably, their community’s expectations. This accusation was the one that initially upset me the most.
The conference's code of conduct (CoC) includes the following clause:
Harassment includes offensive verbal comments related to gender, gender identity and expression, age, sexual orientation, disability, physical appearance, body size, race, ethnicity, religion (or lack thereof), or technology choices.
I can’t help thinking that one of these things is not like the others. Including "technology choices" as a protected characteristic smacks of white techbros trying very hard to find a reason why they are the victim, actually. I find its inclusion deeply offensive, and I question what message it sends about the community's commitment to genuine equity.
To be clear, I’m not suggesting that the CoC should permit harassment on the grounds of technology choices. I feel that the CoC shouldn’t list any specific types of prohibited harassment, because all harassment should be prohibited. The point of listing protected characteristics is to highlight that comments about these characteristics can more easily cross a line into abuse, because they are routinely used as excuses for bigotry.
To assert that “technology choices” belong in the same category as race and gender and similar characteristics requires some kind of conceptual shift regarding the need of marginalized groups for particular protections. It belittles the genuine fear of harassment which so many people face in majority-white, majority-male tech communities.
The conference chair argued that their CoC is a generic document similar to that used by many other conferences. That’s as may be. But too often, “generic” means “written by people in positions of relative privilege and power, without input from those who are most likely to need the protection of a code of conduct”. That clause in the CoC permitted the conference chair to accuse me, formally, of harassment. In my view, it also permitted the community to protect male technologists and their tech choices from ethical scrutiny.
I complained to the conference committee (the chair and two co-chairsThere was a third co-chair, but he’s my husband and therefore quite properly couldn’t be involved in any disciplinary action involving me. about the tone and content of the reprimand. I particularly noted its factual errors, its lack of proportionality or objectivity, and its use of misogynist tropes about how women communicate. I asked the committee to evaluate the process for issuing reprimands, in order to ensure that objectivity and proportionality are at the centre of any official communications, and particularly to pay attention to eliminating systemic biases that punish the speech of groups such as Black people, women, neurodivergent people, and so on.
I got nowhere. I later learned that the conference chair does not involve the conference co-chairs (or anyone else) in any official disciplinary decisions related to the conference. My complaint was therefore about the behaviour of the chair, alone. My complaint would also be evaluated by the chair, alone.
In the context of a professional event like a conference, the conference leadership is in a particularly exalted position of power. In that universe, they have the ultimate power to managing the community’s interactions, thereby shaping the community itself. This is power, but it is also responsibility. The community gathered at this conference has claimed to want to be more inclusive, to encourage diversity, to broaden its scope. Those in positions of leadership therefore have a responsibility to do all they can to achieve this aim.
Would an objective complaints process have found that I needed to be warned for my comments by the conference staff? Maybe. I was uncivil. And if the response had been a quiet warning to watch my tone, I might not even have started to think about the troubling implications of silencing incivility.
But this was not an objective process. It was a hyperbolic, misogynist attack, made with little concern for truth or objectivity, and with the apparent aim of maintaining the amoral facade of civility by silencing uncomfortable truths about the ethical harms of (male) technologists’ choices.
It is impossible for me to feel I belong in a community where a male software engineer is given far greater protection from criticism of his technological choices than I am given from misogyny and ableism; when my incivility in response to a flippant answer is characterized as harassment, yet I have nowhere to turn when I experience ableist comments and misogynist stereotypes directed at me by male attendees and by the conference chair.
When the community accepts a man showing off the use of LLMs to generate alt-text and translations (showing disdain for the needs of disabled and language-minoritized users); when the community accepts another man defending such practice as the best “the disabled” can expect; and yet the same community reacts with outrage and rebuke to a woman who persists in raising the (real, documented, and urgent) ethical issues with LLMs, I think it is reasonable to ask where the moral centre of that community lies.
Since the conference ended, a few fellow attendees have told me that they are grateful to me for raising issues that they feel unable to talk about in this community. I’ll leave it to you to wonder about the demographic characteristics of those other attendees, and about what their silence says about our community.
Towards the end of the conference, in an open discussion, I raised my concerns about “civility fetishism”, which uses a particular, biased conception of civil discourse as a way to dismiss or punish marginalized people when they speak truth to power.
The conference chair responded by proclaiming that they are a proud civility fetishist.
Demands for civility are, ultimately, demands that marginalized voices be silent about their marginalization. Civility has been identified by theorists of race and rhetoric as a trope of white supremacySee e.g. the book by Zamalin in the Further Reading list below.. The fact that civility fetishists are calling for modes of interaction that seem nice, and kind, and balanced, obscures the fact that niceness neuters calls for justice, and balance means doing nothing to redress pre-existing power imbalances.
The truth spoken to our power may transgress against our ingrained sense of what is civil. Our own systemic biases may make us perceive aggression because what we have come to expect is politeness and docility. But we have a duty to hear what is being said, regardless of how it is said. We have a duty to recognise that what is an abstract technological problem for us may harm the real lives of the real people our software touches.
If we are to be a genuinely ethical community, it has to be possible for others to hold us to account, forcefully if need be, when we make harmful choices. Those who hold the most power in the community are also those whose responsibility is greatest, and it must be possible for the community to speak truth to their power.
Our communities require an ever-evolving understanding of ethics and equity if we ever hope to be fit to welcome people who are understandably suspicious of the white, male bias in our demographic makeup. If, instead of resisting being bound by society’s ingrained power imbalances, we permit hegemonic civility to be our straitjacket, then we are condemning our communities to be mere masquerades of monocultural conformity.
Aiston, Jessica (2024) “‘Vicious, vitriolic, hateful and hypocritical’: the Representation of Feminism within the Manosphere” Critical Discourse Studies 21.
Báez, Kristiana L. and Ersula Ore (2018) “The Moral Imperative of Race for Rhetorical Studies: on Civility and Walking-in-White in Academe” Communication and Critical/Cultural Studies 15.
Berenstain, Nora (2020) “‘Civility’ and the Civilizing Project” Philosophical Papers 49.
Cameron, Deborah (2024) Language, Sexism and Misogyny.
Manne, Kate (2017) Down, Girl: the Logic of Misogyny.
Tirrell, Lynne (2019) “Toxic Misogyny and the Limits of Counterspeech” Fordham Law Review 87.
Zamalin, Alex (2021) Against Civility: The Hidden Racism in Our Obsession with Civility.