Introduction
The guidelines for examining patent applications related to artificial intelligence (AI) inventions at the Intellectual Property Office (IPO) are designed to clarify the application of exclusions to patentability and address the sufficiency of disclosure requirements. These guidelines are established under the Patents Act 1977 and the Patents Rules 2007 [2], providing a framework for assessing AI inventions across various technological fields.
Description
Guidelines for examining patent applications related to artificial intelligence (AI) inventions at the Intellectual Property Office (IPO) are established under the Patents Act 1977 and the Patents Rules 2007 [2]. These guidelines clarify how the exclusions to patentability apply to AI inventions and address the requirement for sufficiency of disclosure [2].
Patents are available for AI inventions across all fields of technology [2], provided they meet the four conditions outlined in Section 1(1) of the Act: novelty [2], inventive step [2], industrial applicability [2], and not being excluded by specific provisions [2]. AI inventions are considered computer-implemented inventions that may involve mathematical methods and computer programs [2]. The exclusions apply based on the substance of the invention rather than its form [2], focusing on whether the invention makes a technical contribution to the known art [2].
AI inventions are categorized as either “applied AI,” which applies AI techniques to non-AI fields [2], or “core AI,” which advances the field of AI itself [2]. Both categories may involve training methods [2], such as machine learning [2], and can be implemented in various forms [2], including software [2], hardware [1] [2], or a combination of both [2].
The guidelines emphasize that patent protection is available for training data sets when they contribute technically to an invention [2], but claims solely based on information content are excluded [2]. The sufficiency of disclosure for AI inventions is assessed similarly to other inventions [2], following established legal principles [2].
Examiners are instructed to apply the Aerotel test to determine patentability under Section 1(2) [2], which outlines exclusions for certain subject matter [2], including mathematical methods and computer programs [2]. The test involves evaluating whether the invention makes a relevant technical contribution [2], with five signposts serving as guidance [2].
The assessment of AI inventions requires careful consideration of whether they produce external technical effects or solve technical problems outside the computer [2]. Examiners must evaluate each case on its merits [2], ensuring that the invention’s contribution is not merely the manipulation of data but results in a technical advancement [2].
Recent decisions have indicated that certain AI innovations [2], such as Artificial Neural Networks (ANNs) [1], may face significant challenges in obtaining patent protection under UK law. For instance, a recent ruling determined that an ANN training method qualifies as a computer program under Section 1(2)(c) of the Patents Act 1977 [1], which excludes “a program for a computer” from patentability [1]. This ruling underscores the necessity for AI inventions to demonstrate a real-world technical contribution to be considered patentable.
The guidelines provide a framework for assessing the patentability of AI inventions [2], emphasizing the importance of technical contributions and the careful application of legal principles in the evaluation process [2]. As the landscape of AI patentability continues to evolve, it remains crucial for applicants to ensure that their inventions meet the stringent requirements set forth by the IPO and relevant legal precedents.
Conclusion
The guidelines for AI patent applications at the IPO underscore the importance of demonstrating a technical contribution to the known art. As AI technology continues to advance, these guidelines will play a critical role in shaping the patentability landscape, ensuring that only inventions with genuine technical advancements receive protection. Applicants must navigate these evolving standards to secure patent rights successfully.
References
[1] https://www.eversheds-sutherland.com/en/hungary/insights/patenting-ai-innovations-in-the-uk-part-2
[2] https://www.gov.uk/government/publications/examining-patent-applications-relating-to-artificial-intelligence-ai-inventions/guidelines-for-examining-patent-applications-relating-to-artificial-intelligence-ai–2