Generally, an artificial intelligence (AI) system is a set of algorithms, a mechanism, or a data-analyzing pipeline that is able to discover complex relationships and valuable information from the existing data as well as make decisions and predictions. Today, the success of many corporations, like Google, Amazon, Facebook, and Apple, is beholden to the AI algorithms that enable them to process a huge amount of data collected from different sources, also known as big data. In addition to the aforementioned giant corporations, thousands of fast-growing, successful technology startups have established their businesses based on AI inventions.
In 2019, the World Intellectual Property Organization (WIPO) reported significant growth in the number of AI-related publications and patents. Moreover, it is said that the fourth industrial revolution is happening as a result of AI technology, which has significantly changed business operations and peoples’ lives.
Typically, intellectual property (IP) is the most valuable intangible asset of startup companies; taking this into account begs the question of whether AI-related inventions can be protected under the existing IP laws of the United States; and in particular, whether AI-related claims that direct to the model are patentable. In this blog article, we briefly discuss AI patents as an available, strong method of IP protection.
Patentability of Conventional Software and AI
Under Section 101 of Title 35 of the Patent Act, “whoever invents or discovers any new and useful process, machine, manufacture, … may obtain a patent.” Technically, software, algorithms, and AI algorithms may be patented as a process or a method as prescribed under Section 101 of the Patent Act. Nonetheless, during the many years of examining and interpreting the patent laws, the Supreme Court of the United States has carved out three main categories as unpatentable subject matters: laws of nature, natural phenomena, and abstract ideas. Therefore, although AI algorithms are categorized as a process or a method, they can also fall under these judicially created exceptions as being an abstract idea that is not patentable.
For instance, mathematical formulas or a law of nature like E=mc^2, in the abstract, cannot be patented because it preempts the use of the claimed abstract idea and it gives such power to the patentee to monopolize a field of science. The ultimate goal of the patent system, as stated in the U.S. Constitution, is to “promote the science and useful arts,” and this goal cannot be achieved without making a reasonable balance between granting patents to individuals and protecting the science from being monopolized by granting broad protection to the inventors.
The Supreme Court of the United States introduced a patent eligibility test in its decision in Mayo (2012) and Alice (2014). Under the proposed two-step test, the inventor must first examine whether his claimed invention falls under one of the judicially created exceptions (e.g., abstract ideas), and if so, in the next step, the inventor must be able to show that his claimed invention includes an “inventive concept” that transforms the nature of the claim and amounts to something “significantly more than a patent upon the ineligible subject matter.”
According to the substantial similarity between conventional software and AI algorithms, the legal analysis of software-related patents can also be applied to AI inventions. There are many precedential decisions from the Court of Appeals for the Federal Circuit and the Supreme Court of the United States that are relevant in discussing software patents, but they are out of the scope of this short blog article.
Subsequently, it is notable that due to specific characteristics of AI and software inventions, an inventor must make specific considerations in drafting and prosecuting AI-related inventions. For example, the Court of Appeals for the Federal Circuit, in Enfish (2016), stated that “the software can make non-abstract improvements to computer technology just as hardware improvements can,” and also asserted that an algorithm that “improves the way computers operate” may be a patent-eligible subject matter that is not abstract in the first place. Accordingly, AI inventors may overcome patent eligibility challenges by relying on Enfish and other relevant precedential decisions even without showing an “inventive step,” as may be required under the Alice/Mayo framework.
In this regard and due to the lack of clarity in the laws related to AI inventions, on August 27, 2019, the United States Patent and Trademark Office published a public request requesting comments on the patent eligibility of AI inventions. This public request can be considered as an indication of the lack of clarity in the patent laws concerning AI inventions, and any subsequent changes in the patent laws or the regulations may lead to the invalidity of many existing AI inventions and may affect the outcome of many pending patent applications.
In sum, given the rapid growth in the field of AI inventions and the number of AI-based startups, choosing the right IP strategy is a critical step in the success of AI-based businesses. Patenting can be a strong candidate to protect AI inventions. An AI inventor can utilize a professionally drafted patent as a competitive advantage that enables him to exclude other competitors from making or using his claimed invention.
During this endeavor, and due to the legal and technical complexity of AI patents, working with a renowned law firm in the field of patent prosecution and IP litigation will have a substantial positive impact on the future of business. Michael Ahmadshahi, Ph.D., Esq., Law Offices, with many years of experience in the field of intellectual property and patents, deals with such cases. If you are an AI inventor, entrepreneur, or a business owner, contact Michael Ahmadshahi Law Offices to get initial consultation regarding any potential patent infringement or the right IP strategy to protect your AI inventions.
- See generally Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014).
- See generally Mayo v. Prometheus, 566 U.S. 66 (2012).
- See generally WIPO Technology Trends 2019: Artificial Intelligence, WIPO Geneva: World Intellectual Property Organization (2019), available at https://www.wipo.int/edocs/pubdocs/en/wipo_pub_1055.pdf.
- See generally Enfish, LLC v. Microsoft Corp., 822 F.3d 1327 (Fed. Cir. 2016).