Ecology is not a dirty word

A common argument for ‘ethical’ use of GenAI in academic research and practice is that it’s okay to use these tools to write the final product that you claim authorship for, i.e. drafting, editing or translating sections of a journal article; creating a conference presentation or artistic work; creating a graphical abstract for an original journal article; or writing an outreach article for general audiences.

This is misleading and potentially damaging.

Scholarship is largely defined by original written outputs and there are multiple assumptions underlying a rigorous body of scholarly knowledge:

Academics are Writers. The majority of recognised academic outputs are written texts accessible to multiple audiences.
Writing is an art, not a software prompt or an algorithm-based summary. Text must be created by an author through original creative processes. GenAI tools are not recognised authors. When researchers use them to ‘write’, the tool essentially acts as a scribe for the author’s thoughts. If the author then edits text created by an AI tool, are they still the legitimate author, or are they an editor? Do they waive their IP and authorship rights to the final product?
Expertise grows from an author’s deep intellectual context. An author’s outputs represent their own intellectual environment: background, experience, observations, interpretations, scholarship. This nuance is often essential to interpreting an author’s message and meaning, and it can’t be replicated by an AI tool.
Scholarship grows from diverse voices. This includes diversity of expression. Before the sudden genAI era, one of my joys as a research student supervisor, journal editor and lecturer was mentoring diverse and developing written voices, including different ways to structure sentences, different nuances of language and expression. GenAI has erased this joy. Instead, it supports instant gratification of outputs, rather than development of voice, and it promotes blandness and sameness of written expression. This denies us the diverse discourse we need to progress knowledge and understand context.
Institutions train future scholars. Research students of all levels (PhDs, Masters, Honours, undergrad) are part of the scholarly system and their published outputs contribute to knowledge. Endorsing and encouraging genAI use in teaching, learning and research training ultimately undermines public trust in knowledge.

Claiming originality and integrity

So far, integrity policies from institutions and journals have focused on asking researchers and students to disclose only when they used genAI. This is important, but it unfairly increases visibility to those who choose to use artificial means to progress their careers. Those who choose to rely on their own original voice, imperfections included, are automatically bundled into the ‘unknown’ category.

GenAI has become so ubiquitous and insidious, that uncertainty about AI use now creates ambiguity and silent assumptions about originality. ‘Did they really write all this themselves?’ we wonder, when we don’t see an AI disclosure statement.

It’s no longer enough to assume that genAI wasn’t used if it wasn’t disclosed. Some researchers won’t disclose even when they did use it, and other researchers may use genAI tools for drafting and editing without realising, or while thinking it’s legitimate use that doesn’t require disclosure. GenAI is being forced into so many common workplace software and applications, and institutions and social media increasingly encourage research students to use genAI tools to draft proposals, construct research questions, conduct parts of their literature reviews, and more. This happens despite the vocal efforts of experts and scholars showing how genAI use compromises research integrity, creating confusion and ambiguity. Many students and experienced researchers are unwittingly embedding genAI outputs in their work without considering the implications, often with damaging effects.

Current approaches to disclosure in academic writing implicitly amplify and normalise genAI use, associating those who use it with perceived ‘success’ and sidelining those who prioritise originality. Authorship societies are scaffolding authors and artists with tools and knowledge to claim originality of their work against the genAI barrage, but academia is behind on this.

Academic writers should claim originality

Researchers and students who do not use genAI for any part of their research outputs: claim your work!

Research students who opt-out of genAI use should include a clear statement in their thesis preface to claim the integrity and originality of their work.

Academic writers and publishing researchers should do the same. I’ve added a statement to my website, to publicly claim the integrity and originality of my published work on this site. I don’t use genAI to write, draft or edit any of my blog posts – it’s all me, clumsy wording and typos included. Images are all mine, unless cited as another source (never genAI). I also don’t use it for any of my other outputs, research or general media, and I will include a statement to this effect in future publications whenever I can.

If you’re a researcher, you’re a writer. Protect your work.

Like this:

Like Loading…

Ecology is not a dirty word

Ecology is not a dirty word

Share this:

Like this:

Related

Comments (0)