Can I Protect Something That I Created Using Artificial Intelligence?

The U.S. Copyright Office has also published information about current technologies and their impact on copyrighted works and is conducting a study regarding copyright issues raised by AI. Based on public information through December 2023, the office will evaluate areas for potential congressional action.

With all the buzz surrounding artificial intelligence, common questions keep being asked: what can be protected? And, if I use AI to come up with something new, can I protect the intellectual property? The answers to these questions are complex, but we have some guidance from U.S. Patent & Trademark Office (USTPO). The USPTO has updated its Manual of Patent Examining Procedure to provide inventors and attorneys guidance on what is eligible for protection.

The U.S. Copyright Office has also published information about current technologies and their impact on copyrighted works and is conducting a study regarding copyright issues raised by AI. Based on public information through December 2023, the office will evaluate areas for potential congressional action. This article examines two areas that are not always mutually exclusive: inventions and works of authorship.


Generally, in the United States, a person that invents or discovers any new and useful process, machine, method of manufacture, composition of matter, or any new and useful improvement of these things, may obtain a patent. There are some exceptions. Based upon judicially created law, inventors that attempt to claim rights in abstract ideas, laws of nature and natural phenomena (including products of nature) will be barred from obtaining a patent. According to the U.S. Supreme Court, the reason for this is that laws of nature and natural phenomenon “are the basic tools of scientific and technological work” and that granting a monopoly for this type of discovery would likely impede innovation rather than promote it. These judicially created prohibitions have exceptions, and inventors should consult with a patent attorney or patent agent to determine whether their invention may be eligible for protection.

Generally, there are concepts that the U.S. courts have already determined to be abstract ideas, and thus patent ineligible. For example, certain methods of organizing human activity, such as fundamental economic practices, mathematical formulas, and mental processes. However, under certain circumstances, if the abstract idea contains an “inventive concept” sufficient to “transform” the claimed abstract idea into a patent eligible application, then the inventor may be granted a patent. The abstract idea must include “additional features” to ensure that the claimed invention is more than an effort designed to monopolize the abstract idea. The Supreme Court has explained that merely requiring a generic computer implementation fails to transform an abstract idea into a patent-eligible invention. It stands to reason that AI requires much more than a “generic computer” implementation. The question then becomes whether applying an “abstract idea” or a mathematical formula or algorithm into a complex computer system housing AI would “transform” such into patent eligible subject matter.

Today, we have two important precedents related to artificial intelligence and inventions. First, an inventor must be a human being. In other words, “inventions” generated by artificial intelligence are not eligible for patent protection. Under the patent statutes, the term “inventor” means the individual or individuals who invented or discovered the subject matter of the invention. In 2020, the U.S. patent office rejected a patent application that identified the inventor as DABUS and explained that the invention was generated by AI. The U.S. patent office explained that U.S. patent law does not permit a machine to be named as the inventor in a patent application. The inventor must be a human being.

The second precedent comes from the U.S. patent office concerning an invention for a system and method for improving the transcription of speech into text. The patent office determined that a system that incorporated AI was eligible for patent protection. The invention was directed to a speech recognition system that incorporated AI into the process. In the inventor’s words, the invention included “well-optimized recurrent neural network (RNN) training system that use multiple GPUs, as well as a set of novel data synthesis techniques that allows for a large amount of varied data for training to be efficiently obtained.” Without “geeking out,” a recurrent neural network can be defined as part of what a layperson considers artificial intelligence. The patent examiner had rejected the claimed invention on the grounds that the invention was not eligible for patent protection (among other reasons).

The inventor appealed the rejection, and the Patent Trial and Appeal Board (PTAB) reversed the decision. In its decision, the PTAB explained that for an invention to be eligible for patent protection, the patent office must first look to whether the inventor claims an invention that falls within any judicially created exceptions that are not eligible for patent protection (abstract ideas, mathematical concepts, etc.), and whether the invention includes additional elements that integrate one of the exceptions into a practical application. But only if a claimed invention recites a judicial exception and does not integrate that exception into a practical application, should the patent office undertake additional analysis to determine subject matter eligibility for patent protection. Under such circumstances, then the patent office must determine whether the claimed invention adds a specific limitation beyond the judicial exceptions that is not “wellunderstood, routine, conventional” in the field, or simply appends well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception. Here’s where it gets complex.

The decision delves into the specific steps claimed by the inventor and how the legal analysis should have been applied by the examiner. The important takeaway here is that even if the claimed invention was considered to recite a mathematical concept, under the second prong of the test, the claimed invention was not directed to an “abstract idea” because the judicial exception was integrated into a practical application. Namely, “the claims of the current application include precise features that were specifically designed to achieve an improved technological result” and “provide improvements to that technical field.” The PTAB noted that the claimed invention used DeepSpeech learning, i.e., a trained neural network, along with a language model to “achieve higher performance than traditional methods on hard speech recognition tasks while also being much simpler.” In the end, the claimed invention was using artificial intelligence to improve the conversion of speech into text.

Works of Authorship

Generally, a work of authorship, as defined by U.S. copyright law, includes literary works, musical works (including accompanying words), pictorial, graphic, and sculptural works, motion pictures and other audiovisual works, sound recordings, architectural works, and many other creative works. Even a computer program may be eligible for copyright protection. Copyright protects the expression of an idea that is saved onto some tangible medium—but does not protect the idea itself. In other words, if a person paints a fruit on a canvass, copyright law protects the artist’s creation as the artist expressed it on the canvass, but does not entitle the artist a monopoly on the concept of a painting of a fruit.

The advent of AI, and now, the public accessibility of AI, through free online systems such as ChatGPT, DALL-E 2, Microsoft Bing (with GPT 4), Google Bard, and others, has allowed self-described noncreatives to create incredible “works of authorship.” But is the person engaging AI to create, really an author? The U.S. Copyright Office says, not so fast! The Copyright Office has been receiving and examining applications for registration that claim copyright in AI- generated material. In 2018 the Copyright Office received an application for a visual work that the applicant described as “autonomously created by a computer algorithm running on a machine.” After a series of appeals, the Copyright Office issued a final determination that the work could not be registered because it was made “without any creative contribution from a human actor.”

The U.S. Copyright Office explains that “copyright can protect only material that is the product of human creativity.” The Copyright Office explains that the term “author,” which is used in both the Constitution and the Copyright Act, excludes nonhumans. The Copyright Office’s position on the topic is anchored in an 1884 Supreme Court case addressing photographs, in which the Supreme Court defined an “author” as “he to whom anything owes its origin; originator; maker; one who completes a work of science or literature.” The Copyright Office’s current position is that the Supreme Court referred to such “authors” as human, describing authors as a class of “persons” and a copyright as “the exclusive right of a man to the production of his own genius or intellect.” But this does not mean that AI cannot provide a basis to claim copyright protection in a work of authorship.

The Copyright Office will consider applications for registration of copyright of works that contain human authorship combined with “uncopyrightable material,” including material generated by or with the assistance of AI. The Copyright Office begins its analysis by asking “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work were actually conceived and executed not by man but by a machine.” In its policy guide, the Copyright Office explains: “In the case of works containing AI-generated material, the office will consider whether the AI contributions are the result of ‘mechanical reproduction’ or instead of an author’s ‘own original mental conception, to which [the author] gave visible form.’ The answer will depend on the circumstances, particularly how the AI tool operates and how it was used to create the final
work. This is necessarily a case-by-case inquiry.”

In conclusion, there is no protection for inventions and works of authorship that are purely the creation of AI. However, there is protection available for some inventions and works of authorship that either incorporate AI or utilized AI as an assisting instrument.


(Original article published on