AI and the New York Times: Concerns Rise Over Undisclosed Utilize in Journalism
The line between human and artificial intelligence in content creation is blurring, sparking debate and scrutiny within the media landscape. Recent allegations that a personal essay published in The New York Times’ “Modern Love” column was generated with the aid of AI have ignited a conversation about transparency and the potential impact of AI on journalistic integrity.
The Controversy Surrounding Kate Gilgan’s Essay
The controversy began on March 23, 2026, when Becky Tuch of Lit Mag News raised concerns on X (formerly Twitter) about “I was Deemed Unfit to Be a Mother,” a “Modern Love” essay published in November 2025. Tuch questioned whether the essay, written by Canadian author Kate Gilgan, was created using AI, describing it as “AI slop.” She highlighted the essay’s structure, noting its use of parallelisms – a technique often observed in AI-generated text – as a potential indicator. A specific excerpt Tuch flagged read: “Not hate. Not anger. Just the flat finality of a heart too tired to keep trying.”
Further analysis by AI researcher Tuhin Chakrabarty using Pangram Labs’ AI detection tool estimated that over 60% of the column was AI-generated. However, results varied across different AI detection tools, with some flagging as little as 30% and others finding no evidence of AI use.
Gilgan’s Response and The New York Times’ Stance
Kate Gilgan acknowledged using AI as a tool for “inspiration and guidance and correction” during the writing process, specifically mentioning ChatGPT, Claude, Copilot, Gemini, and Perplexity. She clarified that she used AI as a “collaborative editor and not as a content generator.”
In response to the allegations, a spokesperson for The New York Times pointed to its ethical journalism handbook, which requires freelancers to adhere to established journalistic standards and disclose “substantial use of generative A.I.” The spokesperson stated that journalism at The Times remains “a human endeavor” and that the paper is “consistently assessing best practices” as technology evolves. The Times has not commented on whether Gilgan’s use of AI required disclosure.
A Growing Trend: AI’s Infiltration of Media Outlets
This incident is not isolated. Hachette recently canceled the publication of the novel Shy Girl after readers identified potential AI-generated text. Previously published in the United Kingdom, the novel is now being discontinued there as well. Similarly, the Chicago Sun-Times and The Philadelphia Inquirer were previously found to have published a syndicated summer-reading guide featuring nonexistent novels created using ChatGPT.
Research conducted by Jenna Russell of the University of Maryland and colleagues, using Pangram Labs’ AI detection tool, suggests that AI use is more prevalent in opinion sections of major U.S. Newspapers – including The New York Times, The Wall Street Journal, and The Washington Post – than previously believed.
Challenges in AI Detection and the Require for Transparency
Detecting AI-generated content remains challenging. AI detection tools are still evolving and can produce false positives or fail to identify AI-generated material. Pangram’s CEO, Max Spero, acknowledges the limitations of these tools and the difficulty in accurately estimating the percentage of AI-generated content within a text.
The potential for AI to influence public opinion and policy underscores the importance of transparency. AI-generated language has been shown to be more homogenous and potentially persuasive than human language, and models can be biased towards specific viewpoints.
Moving Forward: Policies, Training, and Watermarking
Addressing this issue requires a multi-faceted approach. Publications can implement clear policies regarding AI use and disclosure, enforce these policies through contracts, and provide editors with training to identify AI-generated text.
Potential solutions as well include government regulations requiring disclosure of AI use and the development of “watermarking” technologies by AI companies to develop it easier to identify AI-generated content. However, OpenAI has expressed concerns about the potential impact of watermarking on user adoption and the possibility of circumvention by “bad actors.”
Related reading