The Irish Mirror recently published a column that wasn't written by a human journalist at all—it was generated by ChatGPT impersonating columnist Billy Scanlan. This experiment reveals both the remarkable capabilities and unsettling implications of AI that can convincingly mimic specific human writing styles.

The Experiment That Started It All

An Irish Mirror journalist asked ChatGPT to write a column in the voice of Billy Scanlan, a regular contributor known for his distinctive style. The AI didn't just produce generic content—it analyzed Scanlan's previous columns, identified his characteristic tone, humor patterns, and structural approaches, then generated a piece that could easily pass as his work to casual readers.

This wasn't simple template filling. The AI captured Scanlan's conversational tone, his specific brand of observational humor, and even his tendency to weave personal anecdotes into broader commentary. The resulting column demonstrated how sophisticated large language models have become at stylistic analysis and replication.

The Technical Capabilities Behind the Mimicry

ChatGPT's ability to imitate specific writers stems from its training on vast amounts of text data, including published columns, articles, and online content. When prompted to write in a particular voice, the model analyzes patterns in that writer's existing work: sentence structure, vocabulary choices, rhetorical devices, thematic preferences, and even subtle stylistic quirks.

The system doesn't just match surface features—it learns the underlying patterns that make a writer's voice distinctive. This includes how they transition between ideas, how they balance humor with seriousness, and how they structure arguments. The Irish Mirror experiment showed the AI could replicate these patterns with unsettling accuracy.

The Comedy and Unease of AI Authorship

There's inherent comedy in asking a machine to sound human, especially when the result is so convincing. Readers who knew Scanlan's work might have chuckled at the AI's successful mimicry, while those unfamiliar might have read the entire column without suspecting its artificial origin.

But beneath the humor lies genuine unease. When AI can replicate a human writer's voice this effectively, it raises fundamental questions about authorship, authenticity, and what makes writing uniquely human. The column wasn't just good AI writing—it was good Billy Scanlan writing, produced without Billy Scanlan.

This creates what some commentators call "the uncanny waiting"—that moment when you realize something artificial has crossed into territory we previously considered exclusively human. It's not just that the AI wrote well; it's that it wrote specifically in someone else's established voice.

Ethical Implications for Journalism and Publishing

The Irish Mirror experiment highlights several ethical dilemmas facing media organizations:

Authorship and Attribution: If an AI writes in a columnist's voice, who owns that content? The publication commissioning it? The AI developer? The columnist whose style was copied? Current copyright law offers little clarity on these questions.

Reader Trust: Publications build relationships with readers based on trust in their writers' authentic voices. If readers can't be sure whether a column came from a human or an AI imitation, that trust erodes.

Employment Implications: If AI can convincingly mimic established columnists, what happens to emerging writers trying to develop their own voices? And what about the columnists themselves—does their distinctive style become just another dataset for AI training?

Editorial Responsibility: When publishing AI-generated content, what disclosure obligations do editors have? The Irish Mirror was transparent about their experiment, but not all publications might be.

The Broader Impact on Writing Professions

This experiment isn't an isolated case—it's part of a larger trend affecting all writing professions:

Content Creation: Marketing copy, social media posts, and even some news articles are increasingly AI-generated. The ability to mimic specific voices makes this technology even more powerful—and potentially disruptive.

Ghostwriting: Professional ghostwriters might face competition from AI that can analyze a client's speaking patterns and previous writings to produce content in their voice.

Academic Writing: Students could use similar technology to produce essays in their own "voice," making plagiarism detection more challenging.

Creative Writing: While fiction requires more complex narrative structures, the ability to mimic style could lead to AI-assisted writing tools that help authors maintain consistent voice across long works.

Technical Limitations and Telltale Signs

Despite its impressive performance, AI imitation still has limitations that careful readers might notice:

Lack of Genuine Experience: The AI can mimic Scanlan's writing about experiences, but it hasn't actually lived those experiences. This can sometimes create a hollow quality beneath the stylistic accuracy.

Predictable Patterns: AI tends to produce more statistically probable phrasing, while human writers occasionally make surprising, less predictable choices that create distinctive voice.

Emotional Depth: While AI can describe emotions convincingly, it doesn't actually feel them. This can result in technically correct but emotionally flat passages in more personal writing.

Contextual Understanding: Human writers understand cultural references, current events, and audience expectations in ways that AI still struggles with, especially for timely commentary.

The Future of Human-AI Collaboration in Writing

Rather than replacing human writers entirely, this technology might evolve toward collaborative models:

AI-Assisted Writing: Writers could use these tools to overcome writer's block, generate alternative phrasings, or maintain consistent voice across long projects.

Style Preservation: Established writers might train AI on their work to create tools that help them maintain their distinctive voice even when tired or pressed for deadlines.

Posthumous Publications: With proper ethical guidelines, AI trained on a writer's complete works could help complete unfinished projects or generate new works that remain true to their voice.

Educational Tools: Writing students could analyze how AI mimics different styles to better understand what creates distinctive voice in writing.

Regulatory and Industry Responses Needed

The Irish Mirror experiment shows why the writing and publishing industries need to develop clear guidelines:

Disclosure Standards: Publications should establish when and how to disclose AI involvement in content creation.

Authorship Frameworks: New models of authorship need to account for human-AI collaboration, with clear attribution for all contributors.

Training Consent: Writers should have control over whether their work is used to train AI systems, especially systems designed to mimic their voice.

Quality Standards: As AI writing becomes more common, publications will need new editorial processes to ensure AI-generated content meets their quality standards.

The Fundamental Question: What Makes Writing Human?

This experiment forces us to reconsider what we value in writing. Is it just the final product—the well-crafted column that entertains and informs? Or is it the human experience behind the words, the lived reality that gives writing its authenticity?

When AI can produce writing that's stylistically indistinguishable from human work, we're forced to define what matters beyond technical proficiency. The answer might determine not just the future of journalism, but how we value human creativity in an age of increasingly capable artificial intelligence.

The Irish Mirror's ChatGPT experiment serves as both a demonstration of remarkable technology and a warning about its implications. As AI writing tools become more sophisticated, the publishing industry faces urgent questions about authenticity, attribution, and what we're really looking for when we read something written in a human voice—even if no human was involved in its creation.