To signal or not to signal? Putting up the anti-AI sign (or not)
By Elizabeth Stice
Alongside the debate about AI use in writing is the debate about how people and publications should indicate their stance. Not all publications have a hard anti-AI stance, but among those who do, some have a notice somewhere on the website, others have adopted a label, such as the “Writers Against AI” mark popularized by Paul Kingsnorth. Just as there are a variety of ways to make peace (or not) with the presence of AI, there are a variety of ways to interpret anti-AI labels and language. Like so many situations in our time, we can approach this through the lens of The Power of the Powerless by Václav Havel.
In The Power of the Powerless, Havel introduces the character of the greengrocer. This theoretical grocer exists in 1970s Czechoslovakia and he keeps a sign in his shop window that reads “Workers of the World Unite.” He does it because it’s what he is supposed to do. He doesn’t necessarily believe in the sign and what it stands for, and, in many ways, he has stopped thinking about the sign at all. If it read something more like “I put up this sign because I am afraid and therefore compliant” he might think much more about it. Havel says this grocer is “living the lie.” He has surrendered his own relationship to reality to ideology. He is succumbing to the dominant ideology and even strengthening it.
It is possible to interpret anti-AI labels as being part of an ideology, which Havel considers a “specious way of relating to the world.” From this angle, some writers and publications have decided that AI is bad and they do not take any context or specifics into consideration. They have a system of ideas that interprets reality for them. They have a sweeping ban without thinking about things like voice transcription or the fact that our search engines have AI and it’s very hard to keep AI entirely out of the writing process in the year of our Lord 2026. There is no nuance. Maybe they even put up the “no AI” sign because there is pressure in the writing community to do that. Maybe they haven’t even thought through what it means. The anti-AI label could even constitute “virtue signaling.” It could be a way of feeling better about oneself and morally superior to those who are less AI-averse.
Yet it is also possible to interpret anti-AI labels as indicators of “small-scale work.” Havel borrowed the notion of small-scale work from T.G. Masaryk and, in Power of the Powerless, Havel defines it as “honest and responsible work in widely different areas of life but within the existing social order.” Havel’s example here is a brewer who was proud of his work and “wanted our brewery to brew good beer.” This brewer was “living in the truth” because he did not see the world through ideology, he saw it as it was and he wanted better beer. His efforts and ideas for improvement shamed and antagonized his superiors who did not especially care about doing a good job and they used the political machine against him.
Then the anti-AI label could be an indicator of “living in the truth.” It can be the sign of a writer who wants to do good, honest work. Or the sign that a publication wants good work, not the kind of cliches that come from AI. The anti-AI label could show that a publication is not just caving to trends or social pressure. Ideology can suggest that AI is inevitable as much as an ideology could suggest that AI is evil. The label, then, has the potential to be offensive to those who embrace AI and feel judged by the label. It could antagonize people the way the honest brewer did. Hostility to the anti-AI label could be grounded in shame.
The perspective on the label may also depend on positioning. Looking out at wider society, where AI use and adoption is widespread and often involuntary, resistance seems to be “living in the truth.” It is harder to avoid AI than to use AI. It is much easier to surrender and be like the greengrocer and go along to get along. Yet within the writing world, there is significantly more pressure to avoid AI than there is in the general public. Writers who admit to using AI are sometimes publicly shamed. And “AI use” is not an especially nuanced category in that world. In those circles, an anti-AI label may seem more like the greengrocer’s sign—you have to put it out to keep good company and stay in business. It is very possible for very reasonable people to look at this situation and have different ideas about who is living in the lie and who is living in the truth and whether or not a website should carry an anti-AI label.
For Havel, part of the problem with ideology is that it sidesteps personal responsibility. Havel writes that ideology “offers human beings the illusion of an identity, of dignity, and of morality while making it easier to part with them. As the repository of something ‘supra-personal’ and objective, it enables people to deceive their conscience and conceal their true position and their inglorious modus vivendi, both from the world and from themselves.” If putting up the anti-AI label is simply following the crowd, it would seem to be offering an illusion of dignity and morality. However, if one is responding to AI use on the basis of their actual morality and from their belief in human dignity, it is hard to argue that putting up the anti-AI label is a shield for a bad conscience or is a specious way of relating to the world.
As we ponder this, we should consider the nature of AI tools themselves and what might be appropriate uses for them. The use of AI tools in writing is not necessarily the same as the use of AI tools in other contexts. Writing is a form of human expression and, arguably, a mode of human thinking. Handing that over to AI tools would seem to be fostering an illusion of autonomy over the real thing. It could even be the height of allowing an intermediary between ourselves and reality. What kind of writer lets tools do the writing?
Havel is also very keen on what he calls the “aims of life.” Havel writes that “life, in its essence, moves toward plurality, diversity, independent self-constitution and self-organization, in short, towards the fulfillment of its own freedom.” In contrast, the “post-totalitarian system demands conformity, uniformity, and discipline.” It would be hard to look at the relationship between AI and writing today and suggest that AI tools are on the side of plurality or independent self-constitution and self-organization. We all know from the troubling emergence of patterns in speech and writing shaped by AI that the trend with AI use is toward conformity and uniformity, if not always discipline. Is the writing produced by AI serving the aims of life? Or serving the aims of some system?
We can go further, by considering the nature of the companies that provide AI tools. Havel suggests another aspect of the aims of life is “the elementary need of human beings to live, to a certain extent at least, in harmony with themselves, that is, to live in a bearable way, not to be humiliated by their superiors and officials, not to be continually watched by the police, to be able to express themselves freely, to find an outlet for their creativity, to enjoy legal security, and so on.” What we know about the current AI companies is that they do not serve the aims of life as Havel describes. They tend toward being continually watched, to having expression muted and homogenized, and, however unwittingly, toward humiliation (or dismissal) by superiors. How could working hand in glove with these companies be considered living within the truth?
On the whole, it seems much less tenable to argue that putting up the anti-AI sign is living in the lie than it is to argue that putting it up could be living in the truth. The anti-AI label is much more than virtue signaling for many people, even if it is virtue signaling for some. Some of those unhappy with the label may well be shamed by the small-scale work of publications that avoid AI, even if others are not shamed and are only dismayed by the lack of nuance.
We return to the perennial question: what is to be done? Havel would counsel us not to hide behind ideology, whether it is for or against AI. Rather, we should take personal responsibility for our use or avoidance of AI. Writers and publications should be able to explain how they believe AI tools should or should not be used in writing and they should disclose that to the public. For some publications the current labels suffice, for others new labels may need to be created. Just as the aims of life do not tend to conformity, we cannot expect all writers to be represented well by a single label even among those who impose limitations on AI use. We can only hope that the disclosures are shorter than CVS receipts and not at all like the rapid-fire speech at the end of medication commercials.
