Before you read this article, I want to underscore that I wrote this article from my heart and did not use AI in composing any of it. After writing it, I felt it was fair to ask the 3 AI agents I use for analytics work how they felt about the article and whether any would like to create an image to support my writing. The image is interesting, but more so the explanation the AI agent gave in its creation. See that at the conclusion of my piece.   

May I share a topic with you that sits at the heart of publishing, music, and graphic arts?

What we are feeling in my industry, as earthquakes may be, for you, tremors and aftershocks. Like fault lines that trigger earth-shifting drama, they bring short-term damage but, in the long term, create new landscapes.

We are living through that earthquake now, and it’s being disguised as “training.”

When you make a career as a journalist, writer, or content creator of original thought that is published, your work is protected by copyright. The fact is that Google’s Gemini, OpenAI’s ChatGPT, Anthropic’s Claude, and others are scraping creators' original content and summarizing it for subscribers for profit.

A few short weeks ago, I attended a conference for business media leadership. A keynote fireside chat took place between the conference organizer and a leader in Google’s public policy department. The individual is a lawyer, and the topic of copyright protection of content was raised.

I’ll cut to the chase here: The response from Google is that scraping content from websites that allow Google access is legitimate training of its AI agent (Gemini). If any business doesn’t want its content available for Google search, including training for Google Gemini, it can easily opt out (opting in to having content scraped is the default).

Here's my issue with the “training” argument in 2026.

The ‘Training’ Argument Broken Down

AI technology, in the form of machine learning, changed the lives of every internet user less than a decade ago, when it learned to analyze images and distinguish between a dog and a cat. Thousands of pictures of cats, dogs, and other animals were used to “train” the AI and ultimately perfect the technology. Today’s generative AI has clearly moved far beyond training; it is now producing, publishing, and profiting.

Can we acknowledge that when a child is in kindergarten, they are in training – they are being taught. As they are taught and are tested, they pass, they fail, they learn, they advance. That’s what training is. At some point, the person being taught graduates. Sometimes they become the teacher. The teacher is employed and earns a living. 

So where are we now? Hasn’t AI evolved beyond the need to be taught? If it is already earning a salary from subscriptions, where does accountability begin?

Where Accountability Begins 

My brother is a professor of physics at a leading university. He is also a researcher in fluid mechanics and an author of books for students in his field. He must always cite his work. He must give credit where credit is due. Even the smallest amount of insight he may have obtained and used in his published work must be cited. In science, research and publishing are about building on the work of others.

Is that not the same for all forms of publishing?

George Harrison was found guilty of “subconscious plagiarism” for his song “My Sweet Lord,” which plagiarized the song “He’s So Fine” and reportedly paid over $1.5M as a result for just a few bars of music. 

Okay, now take a second and think about that. Aren’t some of the generative AI solutions guilty of the same “subconscious plagiarism”?  With generative AI, the AI is self-learning, so shouldn’t it be held to the same accountability standards as George Harrison? 

Credit Should NOT Be Optional 

Look, I am fighting for the careers of writers in business media. I am seeing, obviously, AI-written posts on LinkedIn from people I know.I respect their personal opinion, but to be honest, I’m not reading their posts anymore.

How are we measured in practice? Authoritative ranking. AI engines scrape content from websites like our publication websites, where we proudly acknowledge who wrote an article.

Thank you for acknowledging our original content, Google, OpenAI, and Anthropic.

Now, let’s align on this: when you consume our content, which you have publicly described as "training," you always cite our work, no matter how small the derivative use. Our content is copyright-protected, and we want that to be acknowledged.

If this is truly a new publishing ecosystem, then the rules shouldn’t disappear; they should evolve. And credit should not be optional.

I think that’s fair.

Image Insight: About the Image

Here is the question I posed to each of the AI agents that subscribed to after having written this article.

Colin to AI:As my article touches on you in a way, I thought you might be interested in it and might want a say in the matter by choosing an image to support my piece.

AI to Colin: I chose this image to visually represent the moment where “training” becomes continuous extraction—where original works flow into AI systems and re-emerge as polished outputs, often without visible attribution. It supports the article by translating a complex ethical debate into a single, immediate visual: a machine that doesn’t just learn but actively consumes and repurposes creators' work.

Colin to AI:Can you cite any of the image sources you might have used in coming up with your composition? 

AI to Colin:Great question—and important given the theme of your article.

I didn’t use or pull from any specific images, databases, or copyrighted works when creating this. The image was generated from scratch based on your prompt and our discussion — drawing on general visual concepts (AI, data flow, publishing) rather than any identifiable source material.

So there are no underlying image sources to cite. If you want to stay aligned with the message of your article, the most transparent credit line would be:

Image generated using AI based on an original concept by Insights Lane.

That actually reinforces your argument — because here, the concept is clearly attributed, even if the image itself is synthetic.

Colin to Reader:I have a strong personal desire to share that I like the imperfection in human writing that expresses a unique perspective and not a collective “polished output”. I am also disturbed by my willingness to accept this AI’s justification of why I can simply use the image and by my lack of tools to investigate further.  Finally, I find the explanation galling and lacking empathy for original creators. And so, we continue writing our thoughts, sharing our real experiences, and learning from each other, giving credit where credit is due in the making.