Beyond Human Minds: Exploring the Boundaries of Artificial Intelligence in Patent Law

The journey of Artificial Intelligence (AI) has been very long and fascinating. A number of notable achievements and milestones have marked the evolution of AI. AI is, in some way or another, incorporated into our daily lives. A few decades ago, AI started taking over in those fields when only the human mind was well equipped with the task of playing games, performing complex tasks, researching, or writing. It is now a routine task that these AI-equipped machines are preferred to perform and carry out such complex tasks rather than relying on human skills.[i]  In recent years, AI has attracted significant interest. The pace at which it has come into the mainstream is unmatched. AI chatbots like ChatGPT present a full research paper on one single command. The interface of AI tools is so easy to access that every working professional, including lawyers, finance professionals, researchers, and even students, everyone has become familiar with AI. From virtual assistants, AVs, and chatbots to even medical diagnostics, AI is everywhere.[ii] Its journey has been full of a number of changes and advances, and a lot more is yet to come. It is shaping our lives and will continue to influence how we work and live. The concept of AI dates back to the 1950s when Professor John McCarthy mentioned the term for the first time during the "artmouth Summer Research Project" on AI.[iii]

AI is a complex web of algorithms, innovations, and IPR.[iv] Algorithms are the backbone behind the functioning of AI. These algorithms use a "deep neural network" to take in some principal characteristics. "A deep neural network is a multiple-layered network of inter-connected processors modeled after neurons of the human brain."[v] The Turing test measures artificial intelligence (or "AI"), which is the capacity of machines to react in a manner identical to that of humans. By doing repetitive activities in large quantities, artificial intelligence decreases fatigue. It assists in ignoring human frailties through more thorough data analysis. Due to their extraordinary adaptability and ability to respond to changes in information, AIs continue to evolve to meet new challenges.[vi]

The focus of Al's research seems to be moving from creating machinery that can execute human-like tasks to developing machines with "autonomous, superhuman intelligence," which could threaten existing social institutions based on humans. This is where the Intellectual Property legislation comes to the rescue and serves as the basis for technological innovation. 

Conflicts arising from Al technology's effects on human society initially appear in intellectual property legislation. Al's inventorship problem is complicated and multidimensional. It results from the ability of Al systems to generate discoveries or inventions with little to no human involvement, making it hard to identify the creators of those works.

1. RISE OF AI IN TECHNOLOGY ALONG WITH ETHICAL CONCERNS

Generative AI tools like ChatGPT and Bard are changing industries like education and government, as well as laws and entertainment, with over 100 million users before the beginning of 2023. These tools driving advanced chatbots, AI, "virtual assistants," and "language translation" have multiple features of merits. Misuse, such as copying the work of writers and artists, facilitating cyberattacks, or creating dangerous materials like chemical weapons, is still an ongoing concern.

By spotting intricate patterns, recognizing illnesses early, and refining therapies, machine learning—another expanding AI application—is improving medical diagnostics. Although it has increased access to treatment, especially for underprivileged groups, its efficacy may be diminished by drawbacks such as skewed data, which could result in safety concerns and disparities in healthcare.

A computer uses facial recognition technology to assist crime investigations by instantly recognizing people and matching the travelers to passports. Unfortunately, gross error rates and biases accompany the quest for better accuracy, especially for some groups. Also, there are issues of privacy infringement as well as the wider moral ramifications of its extensive use.[vii]

2. IP LAWS IN OTHER JURISDICTIONS ABOUT AI

Intellectual property (IP) laws give inventors the sole authority to use their creations. For example, patent holders can restrict unlawful use of their "inventions." These safeguards include financial rewards, acknowledging 'artists' accomplishments' while stimulating capital on 'intellectual property.' While this reason has been used in the past to justify IP laws, it becomes immaterial concerning AI –something that creates output that is works of art without needing motivation. This brings up a key issue: should works created by "AI systems" be protected by the same IPR laws as those granted to human creators?[viii]

The EU AI Act aims to protect "democracy, public health, safety, basic rights, and the rule of law" against possible dangers caused by AI while promoting the deployment of trustworthy, human-centered AI technologies. It also aims to guarantee a healthy internal market and encourage innovation. In addition, the AI Liability Directive seeks to address the shortcomings of the existing fault-based liability frameworks in handling claims involving AI-driven goods and services by giving victims of AI-related injury the same degree of protection as those harmed by traditional technology.[ix]

In the United Kingdom, legislation directly addressing the intersection between AI and IP is limited. Meanwhile, Chinese courts have taken a relatively more flexible approach to adjudicating in favor of the copyright on AI-generated content as long as a threshold of human contribution in creating the original work can be established.

3. OVERLAP OF AI IN PATENTS W.R.T. THE DABUS CASE

It is a prerequisite in the patent law that an inventor must be named when any patent application is filed. In the absence of the name of the inventor, the application will be rejected, and the patent will not be granted. Is it possible for an AI system to be acknowledged as a patent inventor?

This was tested when Dr. Stephen Thaler filed a "patent application" in several different applications in several other nations, claiming sole inventor status for "DABUS, an AI system" he owned and developed. The Dabus Case pertains to two UK patent applications, GB1816909.4 and GB1818161.0. Thaler filed these applications under his name with the UKIPO to secure patents for his inventions. In his application, Thaler specified that he was not the inventor of the invention by virtue of Section 30 of the Patents Act, 1977. By replacing reliance on the said section, he asserted that the "right to apply for a patent is transferable." Upon this, the UKIPO informed him and asked him to file the "statement of inventorship" and "right to grant patents" in line with the requirements stipulated under section 13 of the Act. For his part, he filed "Patent Form 7," wherein he went on to state that it was none other but DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), which is an AI device and is the inventor of the current invention as well.[x] Due to Dr. Thaler's ownership status of DABUS, it was assumed that DABUS had given Thaler the authority to grant patents. This assumption raised a plethora of doubts, particularly about whether a patent can be granted to AI.[xi]

Upon hearing, the hearing officer denied recognizing DABUS as the inventor, considering the provisions of the said Act. Furthermore, the ownership of an AI device does not constitute valid grounds for submitting a patent application. His appeals filed in the High Court and Court of Appeal were dismissed.

Appeal to the Supreme Court was concerned with how the 1977 Act's relevant provisions should be interpreted and applied to the patent cases rather than discussing upon the question of whether technological advancements produced by AI-powered, autonomous robots should be eligible for patent protection.

The outcome of the appeal discussed three main points[xii]:

1. Scope and meaning- "inventor" under the 1977 Act

According to the Supreme Court, an inventor under the 1977 Act has to be a natural person, and an AI system like DABUS cannot be one. Section 7(3) of the said act defines the term "inventor as "the actual deviser of the invention." Lord Kitchin pointed out that the ordinary meaning of the term "'deviser' is "a person who devises a new and non-obvious product or process." The right to patent under section 7(2) of the said act is built on the hierarchy wherein it is assumed that "invention" originates from a human creator. This would imply that only entities or corporations that act via an inventor can claim rights in patents, thereby excluding non-human entities, and AI falls within the purview of non-human entities. Further, section 13(1) of the said act ensures that an inventor has the right to be credited for their work. This clearly demonstrates that the inventor is no one else but a natural person. This limits the concept of inventorship to just human beings and, therefore, leaves no scope for a machine or AI system to qualify as an inventor under the said act.

2. Dr. Thaler's patent ownership

As the rightful owner of DABUS, he argued that he had the authority to file for and receive patents for the inventions included in the applications, analogously to or through the application of the theory of accession.

Rejecting Thaler's arguments, the Supreme Court stated that the statute grants derivative rights to non-inventors only when the rights have an original source when they stem from a qualified

Human inventor u/s 7(2)(b)/(c). Contrarily, Dr. Thaler failed to demonstrate any legal basis upon "which he acquired such a right through his ownership of DABUS." Reliance placed on the grounds of the Doctrine of Accession also did not hold any merit in the eyes of the court. The court noted that a tangible property created by another existing tangible property is not the concern here. DABUS is not a legal person, and therefore, the aspect of transferring rights is eliminated here.

3. Hearing Officer's withdrawal decision

According to the Court, the Hearing Officer had the right to conclude that Dr. Thaler did not meet either of the requirements outlined in section 13(2) of the 1977 Act: he did not correctly identify the person or people he thought were the inventors of the inventions listed in the applications, and his ownership of DABUS did not give the Hearing Officer a valid reason to accept his claim that he was entitled to the patents for which he was seeking protection under Patent Laws.

Therefore, the Comptroller was correct in concluding that the petitions would be deemed withdrawn after the period of sixteen months allotted under Rule 10(3) of the Rules had passed. Thaler's patent application has been rejected in other world jurisdictions, namely the United States, Germany, New Zealand, South Korea, and The European Patent Office. Citing similar reasons, a human inventor is a prerequisite of patent laws. Currently, South Africa and Saudi Arabia have accepted the patent application filed by Thaler. Still, the Patent Act of South Africa has laid down various grounds for revoking the application. Time will tell the fate of the patent application in these two jurisdictions.

All in all, since patent rules require a natural person to be recognized as an inventor, almost all of these applications have been denied based on the plain meaning of the language employed in these laws; for instance, according to the European Patent Office, a patent inventor must be human, considering that the details of the inventor must include "family name, given names, and full addresses." Therefore, it is clear that a patent inventor cannot be named after an AI system.

4. WAY FORWARD AND CONCLUSION

The relationship between AI and patents has both advantages and disadvantages. Indeed, the very premise of human inventorship, which has dominated the discourse of science and technology, legal and moral issues up to the present, has been challenged by novel developments in AI invention such as AI invention. There is a fundamental and pervasive tension between encouraging new ideas and innovations and ensuring a reasonable degree of accountability and ethics.

As the laws surrounding AI's role in inventions are still improving and evolving, the invention has multifaceted avenues. For instance, while present laws require humans to be inventors, a mixed approach can be taken where those who build, qualify, and supervise AI systems can be given attribution as inventors. Also, 'AI-Assisted Inventions' and 'AI-Originated Works' categories may be required to fill the voids of conventional intellectual property (IP) systems. Since AI systems cannot be legal persons, the ownership of patents applies to the people or corporations who create, educate, and run them. This would clear up responsibilities and provide accountability for AI-created designs. Given the differences in degree that exist between countries on the AI level and AI patent level, international bodies such as WIPO may play a central role in putting together the guidelines that all countries should follow to tackle the thorny issue of AI inventorship. A more uniform approach would help tackle issues or uncertainty in inventors' patenting and what is ac