The rapid rise of artificial intelligence (AI) has led to significant advancements in various fields, including technological innovation. AI systems are now capable of autonomously generating inventions without direct human intervention. This new reality raises a fundamental question in patent law: Can an invention generated by an AI be protected by a patent?

This complex issue is at the center of concerns for intellectual property offices and legislators worldwide, who are debating how patent law should adapt to this new form of innovation.

 Patent eligibility of AI-Generated inventions

Patentability criteria

Pursuant to the article L611-10 of the French Intellectual Property Code, to be eligible for a patent, an invention must meet three fundamental criteria:

  • Novelty: The invention must not have been disclosed to the public before the patent application was filed.
  • Inventive step: The invention must not be an obvious improvement for a person skilled in the art.
  • Industrial application: The invention must be capable of being used in industry.

When an AI generates an invention, assessing these criteria becomes more complex. The originality of the invention largely depends on the training data and algorithms used by the AI. It then becomes difficult to determine whether the invention is truly novel or merely a reformulation of existing information.

Furthermore, the inventive step requires that an invention is not an obvious outcome of prior knowledge. However, if an AI is programmed to analyze a vast corpus of technical data and propose optimized solutions, can its invention be considered as involving sufficient creative effort?

Challenges in recognizing AI as an inventor

One of the major legal obstacles concerns the attribution of inventorship. Today, most legal frameworks require the inventor to be a natural person.

The European Patent Office (EPO) and the United States Patent and Trademark Office (USPTO) have rejected patent applications where an AI was designated as the inventor.

For example, in January 2020, the EPO rejected two European patent applications in which the designated inventor was an artificial intelligence system named DABUS. This decision was based on the European Patent Convention (EPC), which states that only a human being can be recognized as an inventor.

The applicant, creator of DABUS, argued that this AI, based on artificial neural networks, had autonomously conceived the inventions. However, the EPO concluded that, under the EPC, the rights attached to inventorship, such as the right to be mentioned as an inventor or to assign a patent, can only be granted to natural or legal persons. AI systems, lacking legal personality, cannot be recognized as inventors.

This case highlights the legal challenges posed by artificial intelligence in the field of intellectual property.

These decisions are motivated by the fact that only humans can be legally recognized as inventors, particularly due to legal rights and liability considerations.

This stance raises a dilemma: When human intervention is minimal or nonexistent in the invention process, who should be credited as the inventor?

Some experts suggest attributing inventorship to the AI user or the entity that controls the AI, but this approach remains highly debated.

Legislative developments

International Perspectives

Given the uncertainties surrounding AI and patentability, several initiatives are underway worldwide:

  • The World Intellectual Property Organization (WIPO) has launched consultations on the impact of AI in patent law and is considering potential reforms to harmonize approaches among different countries.
  • The United States has seen the emergence of several legislative proposals aimed at clarifying the legal status of AI-generated inventions, although no major reform has been adopted so far.

In the United States, in October 2023, the White House issued an executive order aimed at ensuring the safe and reliable development of artificial intelligence (AI). In response, the USPTO published, in February 2024, guidelines on the patentability of AI-assisted inventions.

These guidelines specify that:

  • AI may contribute to an invention, but
  • Only a human who has made a significant contribution to each claim can be legally recognized as an inventor.

This position aligns with previous court decisions affirming that the inventor must be a natural person. Thus, patent applications involving AI must explicitly name individuals who have made a substantial contribution to the invention, excluding the possibility of designating the AI itself as the inventor.

This regulatory development highlights the importance of clarifying the respective roles of humans and AI systems in the innovation process, to ensure adequate legal protection for inventions in the United States.

Recent changes in patent law

Some countries have already introduced legislative changes:

  • South Africa became the first country to grant a patent to an AI-generated invention, although this remains an isolated case.
  • Australia also examined the issue, and in a decision on July 30, 2021, the Federal Court admitted this possibility (Thaler v Commissioner of Patents [2021] FCA 879).

These developments show that the recognition of AI-generated inventions is an evolving topic, and regulators will likely need to clarify their stance to address the challenges posed by AI and intellectual property.

Practical Considerations for Innovators

Strategies for Protecting AI-Generated Inventions

Businesses and innovators must anticipate legal challenges by adopting suitable strategies:

  • Ensure a human role in the invention process: A researcher or engineer should be sufficiently involved to be designated as the inventor.
  • Document every stage of creation: Keeping detailed records of how the AI operates and its role in the invention is essential.
  • Explore alternatives to patent protection: When patentability is uncertain, other forms of protection, such as trade secrets, may be considered.

Implications for Intellectual Property management

Companies must adapt their patent management strategies to the challenges posed by AI. It is advisable to:

  • Update contracts to clearly define ownership of AI-generated inventions.
  • Monitor legislative developments to anticipate possible regulatory changes.

Conclusion

The emergence of AI in innovation raises profound legal and ethical questions. Current patent law is not fully adapted to this new reality. Legislators and intellectual property offices must therefore adapt to address the challenges posed by autonomous AI-generated inventions.

Until regulatory clarifications are made, innovators must adopt proactive strategies to effectively protect their inventions and safeguard their intellectual property rights.

Need expert guidance on AI and intellectual property? Dreyfus Law Firm specializes in intellectual property law, including trademark, copyright, patent and AI-related legal matters. Our experts stay ahead of AI and copyright developments!

Dreyfus Law Firm collaborates with a global network of intellectual property attorneys.

Join us on social media !

LinkedIn  

Instagram