Generative artificial intelligence has taken not only the tech world by storm but has also had far-reaching effects on the notions and concepts of intellectual property rights. Ever since the launch of generative AI and the percolation of AI-generated content into mainstream media, there has been considerable debate over the nature of material produced by these artificial intelligence protocols. Since most of the language models are trained on openly available data and content sourced from the internet (of which numerous books, images, and other forms of media belong to their legitimate creators), the furor surrounding generative AI and who owns the content produced by them remains dubious. While there have been numerous cases of AI firms being sued over copyright infringement and other intellectual property rights violations, the debate is still ongoing and there remains no consensus on how to arbitrate these convoluted matters with a factual and measured legal approach.
Image generation models like Midjourney faced copyright violation charges from independent artists and creators for having trained their AI models using privately owned and created pieces of artwork. Similar charges and lawsuits have also been leveled on ChatGPT by authors, who claim the firm used their books to train ChatGPT without their permission. Newspaper outlets like the New York Times even held discussions with top leaders from OpenAI over allegations surrounding the usage of the outlet’s articles for training language models like GPT-3.5 and GPT-4. The famed media house might also consider litigation against the AI firm for the same. Clearly, AI-generated content has thrown the world into a tizzy about copyright law. A few nuances of this phenomenon are discussed below.
Copyright Law and AI-Generated Content
Copyright protection has been offered only to the technologies in themselves and not user-generated images or content.
Copyright law protects original creators from infringement and plagiarism by laying down a proper framework for the content and material they create through a tangible medium. The core purpose of these laws is to protect the progress and creation of scientific and artistic inventions and advancements by offering their creators exclusive privileges and ownership of what they create. This framework might also be extended to organizations and companies, provided there’s tangible evidence of the firm’s employees creating concepts under its ambit. The unique problem with generative AI arises due to the vast data sets most language models use when being trained. Content is often derived from multiple sources, and it is often hard to point to a single origin when a machine learning algorithm creates a piece of art or writes blocks of literary text. AI often relies on utilizing deep learning algorithms to carry out statistical studies, so the protocol can make out tangible connections between information sets and extrapolate based on training data when faced with user prompts. The issue of AI copyright infringement pertains specifically to the dataset, which may contain information sourced from copyrighted content.
While this covers the ownership aspect of AI-generated content, the other part also relates to the use of AI-created material. This is especially important since several digital content creators are rapidly adopting AI into their workflows to enhance the productivity and quality of their output. AI-generated art and images invariably find their way to public forums and platforms where people sharing these pieces of media might have stakes concerning monetary compensation. Currently, there’s little clarity surrounding who really owns the AI-generated content as there are vacillating opinions that stand in favor of both the AI companies as well as the users and prompt engineers. While these matters are being debated by legal professionals, there’s a tangible reason to believe that AI chatbots like ChatGPT have read numerous books and other literary works. A simple prompt requesting the chatbot for a book review or a summary turns up a detailed and mostly accurate response. This creates considerable ground for infringement risk and AI copyright violations.
AI Copyright: What a Legal Approach Might Look Like
Presently there exists little agreement on the scope and extent of planned copyright laws.
Deciding upon a robust framework for creating AI-oriented copyright laws is complex and encompasses numerous stakeholders. Currently, laws in most countries are not oriented toward integrating copyright laws with AI technologies. Present examples include the US statutes where courts have provided copyright protection to only the end product and technology in itself, but not to the user-generated images. While there remains considerable argument in favor of even the latter, courts and legal professionals seem to be studying the implications of such judgments and their effects. Although there have been some regulations concerning the usage of generative AI, there’s no consensus on how content created by AI will be protected, shared, or used for commercial purposes. However, there remains an ever-growing rivalry in the tech markets between AI firms, which is now spiraling into a global technology race. Lawyers and adjudicators will have to make haste in preparing at least a basic outline on how to deal with claimants of infringement as well as the use of AI-generated material in the current tech sphere.
Before moving ahead with enforceable legal frameworks, technological and legal professionals will need to determine the eligibility of AI-created material for copyright protection. Since the act of prompting an AI still does not carry any legal basis for copyright protection as yet, it will be interesting to note how the judiciary manages to create a tangible legal identity for AI prompts and the act of prompting. Given that coding AI platforms like Google Codey and ChatGPT’s Advanced Data Analysis plugin have also rolled out, this phenomenon will go beyond AI writing and AI art and also extend to technological products like computer programs. While the outlook remains muddled, the upcoming years might lend more credence to AI-generated content as tangible legal entities.
The Outlook for AI-Generated Content
It might be a considerable while before tangible AI copyright laws come into existence.
Various companies like Microsoft and OpenAI have pledged to aid fair copyright practices in the AI domain. The US Copyright Office, too, has been involved in efforts to ratify a standard procedural framework to deal with issues concerning generative AI and intellectual property concerns. While most major firms are beginning to work toward bringing about responsible AI practices to their offerings, the issue surrounding intellectual ownership seems to be fairly complex and nuanced. Needless to say, it might take much longer for both the corporations as well as the government to reach a consensus. Legislative efforts are also ongoing, with these concerns being raised in numerous houses of governments across the world. The legal scope of AI copyright and its numerous issues might unfortunately remain in limbo for the time being.
FAQs
1. Can AI-generated content be copyrighted?
Currently, AI-generated content cannot be copyrighted, as a recent judgment from a US court suggested that AI-generated content lacks a sufficient degree of human intervention to be considered an original creation.
2. Can people use AI-generated images commercially?
While individuals can use AI-generated images for commercial use, they only own the individual file and do not possess any copyright claims of the generated image.
3. Do AI copyright laws exist?
Numerous firms, the US Copyright Office, as well as independent governments, are working to create a tangible AI-based copyright framework to prevent future challenges arising from copyright contentions.