“No AI answers, please” – What reporters are really telling sources

• Reporters are screening for AI and saying so plainly in their outreach. Journalists are looking for authentic responses informed by specific expertise and experience.  

• Running your comments through a chatbot to "polish" them can lead to your comments becoming more generic and flagged as AI-written content. 

• Media relations is built on trust, and unfettered use of generative AI in media pitching can erode that trust. 

Recently, we had a client share two versions of a written response to a reporter’s question—one, their own unfiltered words and the other, “polished” by Claude. When they asked us which version we thought would be better to share with the reporter, we recommended that we share their own, original words as the tone and tenor of the original comments were more in line with our client’s true voice—even if they felt less “polished.” 

More and more, we’re seeing reporters add specific notes in their outreach and requests for comment. It’s common now to see a reporter’s note include disclaimers like “no-AI answers please,” “any responses with AI-generated text will not be used,” and even “please share a non-AI headshot.”  

Qwoted, a digital platform that connects journalists with industry experts, recently shed some light on how AI is affecting media relations in its webinar: Slop or Not: Pangram's Max Spero on Authenticity in an AI World. Since December 2025, Qwoted has partnered with AI-detection software Pangram to provide built-in AI detection to both expert sources and reporters who use the platform. 

With how much AI is being integrated into everyday workflows, it’s not surprising that the amount of AI usage in pitching is on the rise—Qwoted found that 80% of media respondents suspect that they receive AI-generated pitches a few times per month and almost 50% encounter them daily or weekly. Yet there is a fine line between using AI to copyedit your comments and asking AI to draft them for you. 

YOU are the expert 

63% of media respondents’ biggest concern when it comes to AI pitching was that AI pitches were more often generic, lower-quality content and wasted their time, according to Qwoted’s survey.  

When it comes to pitching the media, we advise clients to remember that reporters want to talk to them because of their expertise. 

If a journalist wants a generic AI answer, they can open up ChatGPT or Claude in their own browser and type in the questions for their story. Industry experts, on the other hand, bring years of experience from their career and specific, specialized insights based on their expertise—all of these being things that AI cannot replicate. 

For example, a personal finance reporter might want to hear from a financial advisor who is speaking with main street investors on the topics that are top of mind for them. Or another reporter might want to better understand how a portfolio manager for a rising small cap fund is approaching portfolio construction. Both of these examples call for insights from experts with lived expertise. 

The “AI sheen” and how journalists are spotting it 

AI, even when just editing, seems to leave a distinctive sheen on whatever it touches, removing the personal tone of the original author and genericizing the prose. 

Of 54,301 pitches analyzed by Pangram between December 2025 and August 2026, Qwoted found an average AI-likelihood score of 33%, a measure of how much AI-associated writing shows up. That is not to say that more than fifteen thousand pitches were written whole cloth with AI. In fact, many of these flagged pitches were likely “refined,” “edited” or “polished” by an AI chatbot in an effort to make them “sound better.” 

Many journalists follow their intuition to sniff out AI content, with nearly three quarters of media respondents in Qwoted’s survey saying that they rely on their gut instincts to detect AI content. Meanwhile, almost a quarter of journalists are consistently using AI-detection tools.  

Trust is the real currency 

AI detection tools are only becoming more sophisticated. As we shared last week on our blog about AI-watermarking, we’re not sure where the line is drawn between human content edited by AI and AI content edited by a human.  

What we do know is that media relations is built on relationships and trust, and when a journalist starts to feel suspicious that a source used AI to draft their responses, it can be the first step toward eroding that trust.  

The line between human writing and AI polish will keep moving, and detection tools will keep getting sharper. What won't change is the reason a reporter reached out in the first place: they wanted your read on the market, your conversations with clients, your judgment earned over a career.