For Creators, All Is Not Lost

One expert believes creators still have agency and the time to act is now.

Eric R. Burgess, CEO and co-founder of Credtent, argues it’s an American ideal that content creators should be able to consent to how their work is used and be compensated for it. He is convinced, without deliberate and urgent action, it’s a principle that is rapidly eroding in the age of LLMs.

For Burgess, content development and protection is something he knows a lot about. He’s been a creative technologist for around 30 years and has worked with Disney and other big names in designing and producing games and enterprise software.

Together with Dr. Galen Buckwalter, one of the founding scientists of eHarmony, they created Credtent, a set of services to support content creators and help AI companies adhere to responsible practices when it comes to copyrighted material.

They offer an alternative to the legal path, one that is often too expensive for most writers and creatives.

A New Model For Managing AI Content Rights

If a person suspects their material is being used without permission, Credtent will make a request to the AI provider to remove it, recognizing there is complexity here and it may not always be possible to “untrain” the LLM in someone’s content.

Responses and results take time and have been somewhat unpredictable, but early indications suggest some positive movement from the big AI players. Credtent is also now deploying better tools to make it easier for all parties to monitor and manage requests.

One outcome is that a provider will deny that they’ve ingested the copyrighted material in question. Burgess and his team have the technical chops to perform some smart prompting and other approaches that may provide indicative evidence that content has indeed been used in the training of an LLM.

Burgess emphasizes he’s not anti-AI. In fact, he wants AI to succeed for everyone’s advantage. To this end, he offers to value and license copyrighted content to the LLM providers for an annual fee. He sees it as a win-win for creators and the AI industry.

As a recent example, the owners of the content of the Ed Sullivan Show approached Burgess to help determine how much it was worth and how to best monetize it in the era of AI. Done right, everyone benefits, from the content owners, to the LLM and end-users.

If a creator decides not to have their content be part of developing an LLM, he believes AI companies should respect their decision to not be part of the AI economy.

At the center of this debate is a familiar legal question.

The Fair Use Debate

The dominant argument right now for giving the LLMs a free pass to use copyrighted material is that the content they ingest and use meets the criteria for fair use. This means they have the right to use content without permission of its owners based on four factors: 1) the purpose and character of use, 2) the nature of the copyrighted work, 3) the amount of the work copied, and 4) the effect of use on the existing and potential market.

In a legal case, an AI company doesn’t need to prevail in all of the factors to win a fair use determination.

These companies can also maintain that an LLM does not typically produce output that is verbatim of the original content.

In a 2025 case where Anthropic, one of the largest generative AI providers, consumed a vast library of digital books to train their LLM, the judge ruled in their favor citing the fair use doctrine. However, it was a split decision, because the judge also held them liable for illegally downloading millions of pirated books. It was an important case because it demonstrated that it matters both how content is acquired and how it is used, and it showcased the issues ahead as AI copyright cases begin to stack up.

On the topic of fair use, Burgess and his team have a different opinion. They believe fair use is a concept typically associated with educational purposes and when a big tech company processes copyrighted content without consent, sells it and makes money, that’s the definition of unfair use.

In addition, he challenges the assumption that remixed facts can always remain facts. That’s a trust issue and unless resolved, it will continue to be an area of concern for serious uses of generative AI.

Burgess thinks the LLMs that can deliver the most trustworthy content, which, for example, can be validated to be from a human expert, will come out ahead in the generative AI race.

Options For Creators And AI Companies

For creators that wonder whether their content has been used to train an LLM, they should begin their research with their publisher, if applicable. For authors, for example, their publisher may have a clause in their contracts that gives them permission to license their work for AI training.

If content owners want to seek a legal path, they should have a strong argument that unlicensed use of their work is harming its value and is resulting in tangible financial losses. This position is gaining most traction in the court system which will continue to evolve as judges, and the law catches up with the scope, nature, and impact of AI.

Many other options will emerge in the months and years ahead for both creators and AI companies to find amicable resolutions, such as those provided by Burgess and the team at Credtent.

An Unsettled Landscape

For now, AI companies are getting a break from judges as they lean towards a favorable view of fair use. But they must be diligent in how copyrighted works are acquired and used for training their LLMs. In the coming months they may not be so lucky. Law firms are just getting started in their pursuit of major copyright infringements. LLMs may find themselves choosing between licensing and litigation.

These are early days, and the path ahead looks increasingly complex. It may bring a windfall for lawyers, but it will be costly for both creators and AI companies. At the very least, new options are beginning to emerge for those seeking to protect their work.