A conversation I had a few days ago with a colleague has stayed with me much longer than I expected. We were talking about co-authorship integrity and, inevitably, about generative AI. At some point, we discussed an argument that has become increasingly familiar over the past couple of years: that a text produced with the help of AI cannot truly be ours because it does not emerge entirely from our own thinking. Authorship, in this view, depends somehow on the words originating directly from us.<\/p>\n
I have found myself returning to that idea repeatedly. Not because I am convinced that the opposite is true, but because I wonder whether we are asking the question from a rather particular position \u2014 one that many of us, in fact, do not occupy.<\/p>\n
I do not think in English. At least, not entirely. I think mostly in Spanish, although after more than twenty years of reading, writing and researching across languages, I am no longer sure that all the concepts I use belong comfortably to one language or another. Some probably do not. What I certainly do not do is think in the polished, elegant academic English expected by international journals. Yet that is the language in which I am expected to publish if I want to participate fully in many of the international conversations that define my field.<\/p>\n
Like millions of researchers around the world, I have spent my academic career negotiating that distance. For a long time, the accepted solution seemed obvious. I would write as well as I could and then ask a professional translator or language editor to help produce a manuscript that met the linguistic expectations of international publishing. Sometimes journals explicitly recommended it; occasionally, they even required evidence of professional language editing. Nobody ever suggested that this somehow weakened my authorship. If anything, it was considered good academic practice.<\/p>\n
And this is where I have started to feel uncomfortable with some of our current conversations about AI, because translation has never been a neutral operation. Neither has language editing. A translator does not simply replace one word with another, and an editor does not merely correct commas and verb tenses. Choices are made. Sentences are reorganised. Ambiguities disappear or emerge. Some formulations become more assertive; others more cautious. Concepts move, sometimes slightly and sometimes considerably, as they cross languages and intellectual traditions.<\/p>\n
Educational Technology may be an especially interesting place from which to think about this. Our field has spent decades importing concepts across languages while often assuming that translation was merely linguistic rather than epistemological. Ours is a field that has travelled internationally through concepts that often appeared to be equivalent when they were not. Pedagogy<\/em> became pedagog\u00eda<\/em>. Education<\/em> became educaci\u00f3n<\/em>. Didactic<\/em> was read through traditions of did\u00e1ctica<\/em>. But these are not equivalent terms. Pedagogy<\/em>, pedagog\u00eda<\/em>, education<\/em>, educaci\u00f3n<\/em>, didactic<\/em> and did\u00e1ctica<\/em> do not simply name the same things in different languages; they belong to different intellectual traditions and carry different assumptions about teaching, learning, knowledge and educational practice. When we translate them as if they were interchangeable, we do not merely lose nuance. Sometimes, we reorganise the problem itself.<\/p>\n I sometimes wonder how many disagreements in our field have actually been disagreements between conceptual traditions that happened to use similar words. And yet we have rarely treated these mediations as a threat to authorship. Nor have we seriously questioned whether ideas somehow ceased to belong to those who produced them simply because they had travelled through translators, editors, reviewers, collaborators or different linguistic traditions.<\/p>\n This is probably why I struggle with the way we currently frame conversations about AI. When I use AI while writing in English, I am not usually asking it to think on my behalf. More often than not, I am trying to preserve a thought that already exists as it moves into a language that is not entirely my own. Sometimes the AI finds a formulation that captures exactly what I meant; sometimes it misses the point completely; sometimes it produces something smoother but conceptually worse. And sometimes, rejecting its suggestions helps me understand my own argument more clearly.<\/p>\n The thinking is still mine, but the negotiation has become visible. Perhaps that is one of the things that has really changed. For many researchers who work across languages, mediation is not new; it has always been part of academic writing. What is new is that we are beginning to notice it \u2014 or perhaps that a particular kind of mediation has become visible enough to trouble us.<\/p>\n At the same time, I find it difficult to ignore a contradiction. Many journals now accept, or at least tolerate, the use of AI for grammar, fluency and style. Apparently, that kind of mediation does not fundamentally trouble our understanding of authorship. Yet the moment the conversation with AI begins to influence how an idea is articulated, we become uneasy. I am genuinely not sure where that boundary lies, nor am I convinced that grammar, style and thinking have ever been as separable as we sometimes pretend.<\/p>\n Anyone who has tried to move a difficult idea from one language into another knows this. Sometimes you discover that you do not have the exact word you need. Sometimes the available word carries a tradition you did not intend. Sometimes changing the sentence changes the argument. Sometimes finding a formulation is part of finding out what you actually think.<\/p>\n This is why I increasingly feel that this is not only a conversation about AI. It is also a conversation about who has historically been able to participate in academic knowledge production without having to negotiate another language; about whose ways of writing have become the norm against which everyone else is measured; about who can write directly into the dominant language of international academia and who must pay \u2014 with money, time, dependence, uncertainty, or all of them \u2014 for mediation. And it is a conversation about the forms of mediation we have normalised for decades without considering that they might reflect asymmetries of power, perhaps even a subtle form of academic and linguistic colonialism.<\/p>\n I use the word perhaps<\/em> deliberately. I am not sure I have fully worked this out yet. But I find it increasingly difficult to discuss AI, writing and authorship without also asking why some forms of mediation have historically appeared natural, professional and intellectually harmless, while others suddenly provoke anxiety about whether the words \u2014 and therefore the ideas \u2014 are really ours. Maybe the problem is not simply that AI mediates. Maybe the problem is that AI has made mediation difficult to ignore.<\/p>\n So I find myself wondering whether we are asking the wrong question. Rather than asking whether AI makes a text less \u201cours\u201d, perhaps we should be asking what authorship has actually meant all along. Has it ever depended on producing every sentence in complete isolation? I doubt it.<\/p>\n Academic writing has always been mediated: by language, by disciplinary conventions, by reviewers, editors, translators, colleagues, supervisors, co-authors, institutions and technologies. That does not mean that every form of mediation is equivalent. Clearly, it is not. Nor does it mean that questions of integrity disappear. Quite the opposite. We need much better questions about dependency, substitution, opacity, responsibility and the point at which assistance becomes something else.<\/p>\n But perhaps authorship has never depended primarily on the absence of mediation. Perhaps it depends on something more demanding: understanding the ideas we put into the world, being able to justify them, recognising the intellectual traditions and contributions on which they depend, and ultimately taking responsibility for them.<\/p>\n I certainly do not have a definitive answer. But I have the feeling that the conversation we need to have is much larger than AI itself. Perhaps AI has not disrupted authorship nearly as much as it has disrupted our ability to ignore the mediations that have always shaped academic knowledge.<\/p>\n Inspired by ongoing conversations around co-authorship integrity, and particularly by:<\/em><\/p>\n Ebrahimzadeh, M., Shibani, A., & Buckingham Shum, S. (2026). Coauthorship integrity: Reconceptualising assessment validity for the age of generative artificial intelligence. Computers and Education: Artificial Intelligence, 10<\/em>, 100609. https:\/\/doi.org\/10.1016\/j.caeai.2026.100609<\/a><\/p>","protected":false},"excerpt":{"rendered":" A conversation I had a few days ago with a colleague has stayed with me much longer than I expected. We were talking about co-authorship integrity and, inevitably, about generative AI. At some point, we discussed an argument that has become increasingly familiar over the past couple of years: that a text produced with the Read More<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"episode_type":"","audio_file":"","podmotor_file_id":"","podmotor_episode_id":"","cover_image":"","cover_image_id":"","duration":"","filesize":"","filesize_raw":"","date_recorded":"","explicit":"","block":"","itunes_episode_number":"","itunes_title":"","itunes_season_number":"","itunes_episode_type":"","_vp_format_video_url":"","_vp_image_focal_point":[],"footnotes":""},"categories":[27],"tags":[],"class_list":["post-2704","post","type-post","status-publish","format-standard","hentry","category-trabajos"],"_links":{"self":[{"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/posts\/2704","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/comments?post=2704"}],"version-history":[{"count":3,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/posts\/2704\/revisions"}],"predecessor-version":[{"id":2713,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/posts\/2704\/revisions\/2713"}],"wp:attachment":[{"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/media?parent=2704"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/categories?post=2704"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.lindacastaneda.com\/en\/wp-json\/wp\/v2\/tags?post=2704"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\n