When AI Writes Your Byline, Readers Deserve to Know
An academic's opinion piece in a major Australian newspaper triggered a backlash when readers discovered it was written by artificial intelligence-and the newspaper hadn't disclosed that fact.
The University of Western Sydney's pro vice-chancellor for quality and integrity published a rebuttal to criticism about AI use in student essays. The piece itself was generated by Copilot, Microsoft's AI assistant. The university defended the practice, saying the AI was used "with" the author, not "by" her, and that it helped with drafting, structure, and language refinement.
That distinction misses the point.
The reader's right to know
Opinion writing is not a neutral delivery of facts. It's an exercise in persuasion. Readers engage with an opinion piece to weigh the writer's insights, appreciate their prose, and consider their specific arguments. The author's voice and judgment carry weight in shaping how readers think.
Every reader deserves advance notice of whether they're reading words crafted by the named author or generated by a machine. Some readers may accept AI-written pieces. Others may decline. That choice belongs to them, not the publisher.
Where the line should be drawn
AI can handle legitimate support tasks: researching facts, testing ideas, checking spelling and grammar, formatting tables and bibliographies. These are tools, like a dictionary or spell-checker.
AI must not write the sentences and paragraphs themselves. That's the boundary.
These limits are what some call de minimis standards-specific rules that leave no room for interpretation about what's acceptable and what isn't. Statements of principle alone don't work. People disagree about their meaning.
The slippery slope
If one publication runs an AI-written opinion piece without disclosure, others will follow. Someone might respond with an AI-generated rebuttal. Soon the norm shifts. Human authorship becomes irrelevant rather than the default.
The Sydney Morning Herald and The Age responded by removing the piece and pledging to ask new contributors to guarantee that AI didn't write or construct their articles. Both outlets should publish their specific de minimis standards publicly.
If contributors violate those standards, publishers should implement technological verification of human authorship.
Universities lag behind
Universities are working to integrate AI into teaching while protecting learning outcomes. But they've moved slowly compared to how fast students have adopted AI tools.
The Castlereagh statement issued this year by most Australian universities and educational associations acknowledges how AI affects teaching and learning. It discusses preparing students for an AI-transformed future. It proposes no specific rules or timelines.
Concrete policies are overdue. Students and academics should produce work that is unquestionably their own-not subcontracted to machines.
For writers, the stakes are direct. Your byline signals accountability and authorship. Readers trust that signal. Protecting it protects your credibility and theirs.
Learn more: AI for Writers and Generative AI and LLM resources can help you understand these tools and how to use them responsibly.
Your membership also unlocks:
