Are computer-implemented inventions entitled to be protected by intellectual property law?

The patentability of computer-implemented inventions (CIIs), particularly software, remains a contentious and complex issue within intellectual property law. The controversies are characterised by persistent legal uncertainty and divergent jurisprudential approaches across major jurisdictions like the US and Europe. This thesis investigates the essential question of whether, and under what conditions, computer-implemented inventions are entitled to be protected by intellectual property law; primarily with a focus on the patent protection for software. It critically examines these issues based on the context of the dynamic smartphone industry whilst also acknowledging the emerging challenges posed by Artificial Intelligence (AI). This thesis investigates the essential question of whether, and under what conditions, computer-implemented inventions are entitled to be protected by intellectual property law. In doing so, it considers patent protection not merely as a reward for R&D investment, but as a critical mechanism for defining proprietary boundaries in the market, thereby ensuring that intangible assets can be securely traded and commercialised. This study identifies that while copyright offers foundational protection for the literal expression of software code against literal copying (e.g., 'cut and paste' infringement), as mandated by international agreements like the TRIPS Agreement implemented in regional and national laws, it inherently fails to demonstrate same precise protection to the functional aspects that often constitute the core innovative value of software. Patent law, in principle, is essential for protecting the technical inventions and functional concepts inherent in the software. However, the application of traditional patentability criteria—novelty, inventive step (non-obviousness), and industrial applicability—to software is fraught with difficulties. These challenges are exacerbated by statutory exclusions for "programs for computers as such" (under the European Patent Convention) or judicial exceptions for "abstract ideas" (following the US Alice Corp. decision), leading to inconsistent examination practices and high rates of patent invalidation in courts (as evidenced in Chapter 5 and 6). Furthermore, this study clarifies that while copyright acts as a fallback for protecting literal code, it fails to protect the underlying functional innovation. This study employs the patent practices in the US and Europe. It scrutinises historical evolution, from early industrial eras to the current knowledge-based and AI-driven economy, with landmark case law and patent office guidelines to identify systemic problems including legal uncertainty, patent quality concerns, impacts on innovation incentives, and competition issues. A timely analysis of the profound implications of AI and generative AI for software is also introduced. Key findings show that unclear and inconsistent criteria for software patentability create legal uncertainty and market distortions. The study finds that while both US and EU systems allow software patents under certain conditions, their tests and rationales differ, causing challenges for global innovators. The thesis argues that neither a full ban nor an overly permissive approach effectively serves the patent system's goals for software. To address these deficiencies, the thesis proposes a novel, structured five-dimensional examination model, as a sui generis examination model applied within the existing legal frameworks, rather than advocating for a new legislative act for software patents, which integrates five core evaluative aspects: novelty, inventive step (non-obviousness), technical character (technical effect), industrial applicability, and economic significance. Each aspect is further delineated by specific indicators designed to enhance consistency, predictability, and the quality of patent examination for computer-implemented inventions. The application of this model to leading patent cases demonstrates its potential utility in providing a more transparent and doctrinally grounded framework for assessing software-related inventions. Ultimately, this thesis contributes to the existing body of knowledge through the creation of the abovementioned model which aims to assist patent offices, applicants, and judicial bodies in navigating the complexities of software IP, further fostering a more balanced and effective patent system that supports genuine technological innovation in the software industry.