As artificial intelligence continues to revolutionize content creation, questions surrounding intellectual property rights have become increasingly complex. One of the key dilemmas centers around ownership: Who rightfully possesses the content generated by AI systems? Is it the user who inputs the initial data and parameters, or does ownership lie with the AI itself, as the creator of the output?
In traditional content creation, ownership is typically clear-cut: The creator of the work holds the intellectual property rights. However, with AI-generated content, this delineation becomes blurred. While users may input data and parameters into the AI system, it’s the AI algorithm that processes this information and generates the output autonomously. As a result, determining ownership becomes a nuanced legal and ethical issue.
Some argue that, since the user provides the input data and parameters, they should retain ownership of the content generated by the AI. Others contend that, since the AI system autonomously generates the output based on its algorithms and learning mechanisms, it should be considered the creator and thus hold the intellectual property rights. Additionally, there are concerns about potential infringement if AI-generated content incorporates copyrighted material without proper authorization.
The need for clear legal frameworks and guidelines regarding AI-generated content ownership is evident. Without established regulations, disputes over intellectual property rights could escalate, leading to legal challenges and uncertainties. As the capabilities of AI continue to evolve, policymakers, legal experts and industry stakeholders must collaborate to develop robust frameworks that address these complex issues and ensure fair and equitable treatment for all parties involved. Furthermore, proper citation and acknowledgment of AI-generated content sources are essential to maintain transparency, integrity and respect for intellectual property rights in the digital age.
I hope you liked that Introduction. I didn’t write a word of it. And we don’t claim rights of copyright in that text. We generated a request to a generative AI platform and — voila! — it spit out this wonderful text to get us into the discussion of AI.
What’s factually wrong and/or out of date:
It is incomplete and perhaps misleading: There are many “key dilemmas” that AI presents and focusing only on IP — which is very important — is only one. I will address some of them below.
Ownership of content is not clear-cut: There are many factors going into the issue. Only one example is content created by a true independent contractor. It is the contractor who owns the content unless the contractor assigns all rights of copyright to the owner. Determining who is a contractor and who is an employee working within the scope of the employee’s duties to their employer is also a somewhat difficult question, and one often overlooked.
The U.S. Patent and Trademark Office has already chimed in and has announced that computer generated inventions (patent) are not owned as we humans think.
More importantly for content, the Library of Congress (which oversees the Copyright Act) has announced that content generated by AI is not owned by the person who input a request and received the AI-generated content in return. The Copyright Office previously has ruled that a photograph taken by a monkey is not copyrightable subject matter, for example — and monkeys are living beings.
In other words, the algorithms that generated AI content cannot own IP rights, as least as the issue stands now. The argument will go on, I am sure.
Disputes over AI-generated content are already proliferating.
Celebrities are suing over “deep fakes” of their likeness, some used in a pornographic manner.
Photographers and artists are doing the same; they argue that AI’s “scraping” of content — their IP — is an infringement of their rights. Major publications like the New York Times are heavily involved in suing AI engines for scraping NYT content to “teach” the AI algorithm and thus allow NYT content to be used in generating AI inputs.
Class actions have been filed against major technology companies by authors claiming that the ingestion and unauthorized copying of their content infringes their rights of copyright.
The AI-generated introduction to this article fails to mention rights of publicity, which recently has been renamed as “NIL” — name, image and likeness rights. A photo is a copyrightable item, but its depiction of a human being and the later use of that photo may infringe NIL rights — something that is rarely recognized as a right separate from copyright.
There’s more.
“Hallucinations” has a new meaning in the AI world. It means false information that AI generates. Only one example: A lawyer in New York used AI to generate a brief that included several case citations to support his argument. Only the cases the AI generated never existed. He got off with a small fine of $5,000. The lesson: Fact-check everything that AI generates before you use it.
The AI-generated introduction to this article also fails to mention the risk of false advertising, closely related to the risk of hallucinations. Consider: If you write articles, post content about your products and services or make other use of AI-generated content without checking it first, you may be liable for false advertising under federal and state laws.
The next issue: What happens to your content? Does the AI engine grab it and use it to teach its algorithm? Is the AI engine infringing your rights? Some proponents of AI are arguing that this teaching process is fair use, but, in this author’s opinion, that’s a stretch. As the old commercial once admonished: It’s 10 p.m., do you know where your children are?
Legislation is coming. Bills have been introduced in Congress to regulate the use of AI. Foreign countries — especially the European Union — have already drafted (and, perhaps, enacted) guidelines for the use of AI. So, the rapidly evolving environment of AI will include new laws and regulations requiring compliance at the risk of civil liability and, perhaps, criminal liability in the worst cases.
The lessons: Be afraid, dear reader. Be very afraid.
And, of course, you need competent legal counsel to help you through this quickly evolving muck. Why do you think lawyers write these updates?
William Heller is of counsel with Genova Burns in Newark.
