Did ChatGPT Really Copy Me? OpenAI Pushes Back on New Copyright Claims by Authors Guild
The legal battle between OpenAI and a group of prominent authors, including John Grisham, George R.R. Martin, and Jodi Picoult, has taken another turn. In a recent court filing, OpenAI is asking a federal judge to scale back a new line of copyright claims brought by the Authors Guild, arguing that these broadened allegations stretch legal logic and copyright law too far.
At the heart of the dispute is whether the outputs of ChatGPT, OpenAI's popular AI tool, can be considered direct copyright infringement. Originally, the lawsuit focused on OpenAI's use of copyright works to train its large language models. Now, the plaintiffs have introduced a new theory: that ChatGPT's responses themselves could violate authors' copyrights if they are "derived" from protected texts.
In the 15 July filing in a Manhattan federal court, OpenAI called the updated complaint "radically changed" and "much broader." The company argues that the plaintiffs have not provided any real examples showing how ChatGPT's outputs actually infringe on their specific works. In OpenAI's words, there is no "meaningful description of what the outputs contain."
That lack of detail could prove crucial. Copyright law is not just about inspiration, it is about copying by derivation. So far, OpenAI says there is no clear allegation that ChatGPT is reproducing significant, recognisable portions of any author's copyright work.
The Stakes for Authors
For writers and publishers watching this case unfold, the implications are enormous. The first wave of lawsuits, filed in 2023, entered on whether using copyright works to train AI models, without permission, infringed copyright law. Now, with the Authors Guild introducing the idea that the outputs of those models are infringing, the debate is shifting into even murkier territory.
If the court agrees with the Guild's new theory, it could reshape how generative AI tools are developed and used, potentially opening the door to thousands of similar claims by authors, journalists, and content creators.
However, if OpenAI succeeds in getting this broader claim dismissed, it would reaffirm a narrower view: that using publicly available texts to train AI is legal, and that outputs are only infringing if they clearly copy original expression.
Why It Matters Now
This legal battle is more than just a tech-versus-creatives standoff. It is about where we draw the line between innovation and intellectual property. For authors, the fear is simple: if AI tools trained on your books can reproduce stories that echo your style, or worse, mimic your words, what happens to your rights? On the other side, OpenAI and other developers argue that AI models do not "remember" or reproduce works in the way humans might. Instead, they generate new content based on patters learned from huge datasets, including copyright materials, but also countless other sources.
What Happens Next?
The judge in this case will now decide whether the Authors Guild's new claims about ChatGPT's outputs can stand. That decision could help set the rules for what AI companies must do to avoid legal risk and what authors must show to prove they have a copyright claim.
In the meantime, this case is a wake-up call for writers: AI is here and it is re-writing the boundaries of copyright law in real time. Whether you are concerned about your works being used for training, or the possibility of AI echoing your voice too closely, it is clear that authors need to stay engaged in this debate.
Make an Enquiry Now
For advice on how best to protect your work as an author or publisher, speak to one of our intellectual property experts on 0131 478 4724 or complete an Online Enquiry.
We have helped hundreds of individuals and businesses.
