AI Is Capable of Being Patented, UK Supreme Court Rules.

Introduction and Overview of Recent Developments

The United Kingdom Supreme Court has recently addressed two fundamental questions concerning artificial intelligence and patent law, delivering judgments that significantly shape the legal framework governing AI-related intellectual property in the United Kingdom.

These decisions concern, first, whether artificial intelligence systems may be named as inventors on patent applications and, second, the patentability of AI-implemented inventions themselves.

On 11 February 2026, the Supreme Court delivered a landmark judgment in Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks [2026] UKSC 3, ruling in favor of the appellant and significantly clarifying and expanding the scope of patentability for artificial intelligence-related inventions under United Kingdom law. This decision must be considered alongside the Court's earlier unanimous ruling in December 2023 in Thaler v Comptroller-General of Patents, Designs and Trade Marks, which definitively established that artificial intelligence systems cannot be named as inventors on patent applications.

Together, these judgments establish the current state of United Kingdom law on artificial intelligence and patents; AI-implemented systems may themselves be patentable subject matter, but only natural persons may be designated as inventors. This creates a framework that protects AI-driven innovations while maintaining the traditional requirement of human inventorship.

Introduction and Factual Background of The Emotional Perception AI Case

A.I can be patented full case.pdf

Download PDF • 302KB

On 11 February 2026, the United Kingdom Supreme Court delivered a landmark judgment in the matter of Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks [2026] UKSC 3. The Court ruled in favor of the appellant, Emotional Perception AI Ltd, thereby significantly clarifying and expanding the scope of patentability for artificial intelligence-related inventions under United Kingdom law, which implements Article 52 of the European Patent Convention.

The case concerned Emotional Perception AI's patent application for a system employing an artificial neural network trained via machine learning techniques.

The invention analyzes users' emotional responses to media content, including music and films, and subsequently recommends content likely to evoke similar emotional reactions. The system's recommendations are based upon both emotional similarity and musical properties, thereby extending beyond conventional genre-based or taste-based recommendation engines.

The applicant asserted that this approach delivered faster, more accurate, and superior results compared to existing technologies.

The United Kingdom Intellectual Property Office initially refused the application in June 2022, classifying the invention as a "program for a computer…as such" under section 1(2)(c) of the Patents Act 1977, a category statutorily excluded from patentability. This refusal precipitated a series of appeals through successive judicial levels.

Procedural History

Emotional Perception AI Ltd, which had originally filed the application under the name Mashtraxx Limited in April 2019, sought patent protection for a system utilizing an artificial neural network trained through machine learning. The invention measures emotional similarity beyond mere genre or taste preferences, alongside objective musical or file properties, to recommend new content calculated to evoke comparable emotional reactions.

The applicant contended that the system demonstrated superior speed, accuracy, and relevance when compared to existing recommendation engines.

Following the UKIPO's refusal in June 2022, the matter proceeded through a multi-stage appellate process. The High Court, Patents Court, under Sir Anthony Mann in November 2023, overturned the UKIPO's refusal, holding that an artificial neural network does not constitute a computer program at all. However, the Court of Appeal reversed this determination in 2024, reinstating the exclusion and upholding the UKIPO's original decision.

Appeal to the Supreme Court

Emotional Perception AI Ltd appealed to the Supreme Court, where the matter was heard in July 2025. Judgment was handed down on 11 February 2026. The Supreme Court panel comprised Lord Briggs, who delivered the unanimous judgment, with Lords Hamblen, Leggatt, Stephens, and Kitchin concurring.

Key Holdings and Legal Reasoning

The Supreme Court's unanimous judgment established several crucial principles that fundamentally reshape the legal landscape for patent eligibility in the United Kingdom.

Abandonment of the Aerotel Test

The Court explicitly rejected the four-step "contribution" approach established in Aerotel Ltd v Telco Holdings Ltd [2006] EWCA Civ 1371, which had dominated United Kingdom patent practice for nearly twenty years. The justices ruled that Aerotel misinterpreted Article 52 of the European Patent Convention by improperly conflating the threshold question of whether something qualifies as an "invention" with subsequent assessments of novelty, inventive step, and industrial application. The Court endorsed criticisms of this approach articulated by the European Patent Office, including those expressed in T 154/04 Duns Licensing Associates and impliedly in G 1/19.

Adoption of European Patent Office Practice

The Court held that United Kingdom courts and the UKIPO must now align with the European Patent Office's framework, particularly the Enlarged Board of Appeal decision in G 1/19 (Pedestrian Simulation), decided in 2021. This framework incorporates the "any hardware" principle: any claim involving physical hardware elements confers technical character, thereby satisfying the initial eligibility threshold under Article 52 of the European Patent Convention. The Court characterized this as a "very low hurdle." Following this initial determination, subsequent assessment employs the structured Comvik approach, as articulated in T 641/00, to distinguish technical features from non-technical features when evaluating inventive step.

Classification of Artificial Neural Networks

Contrary to the High Court's determination, the Supreme Court held that artificial neural networks constitute "programs for a computer" in a technologically neutral sense. Their architecture, including weights, biases, structure, and activation functions, operates as instructions executed by computing devices, whether implemented through software or hardware.

Application of the "As Such" Exclusion

The Court emphasized that the statutory exclusion for computer programs applies only to programs "as such." Because artificial neural networks inherently require physical computer hardware to operate and undergo training, even commonplace hardware suffices for this purpose, inventions involving such networks possess technical character and therefore qualify as patentable subject matter. The Court determined that the specific claims contained within Emotional Perception AI's application cleared this threshold.

The Supreme Court remitted the application to the UKIPO Hearing Officer for reassessment under the newly adopted European Patent Office-aligned framework. This reassessment must address whether the invention meets the requirements for demonstrating a "further technical effect," as well as satisfying the criteria of novelty and inventive step.

Broader Implications for Patent Practice

This ruling is widely regarded throughout the legal and technology communities as strongly pro-innovation and represents a major step toward harmonizing United Kingdom practice with that of the European Patent Office. The decision reduces the divergence between United Kingdom and European Patent Office approaches that had persisted since the European Patent Convention's implementation in domestic law.

Legal practitioners and commentators have characterized the judgment as "seismic," a "game-changer," and a "boost for innovation," particularly significant given the current expansion of artificial intelligence technologies. The ruling lowers the initial subject-matter eligibility threshold for numerous artificial intelligence, machine learning, and software inventions that involve hardware implementation. This development may encourage increased patent filings in the United Kingdom and provide stronger intellectual property protection for domestic technology companies.

The Thaler Case: AI as Inventor

In contrast to Emotional Perception AI, which concerned the patentability of AI-implemented inventions, the earlier case of Thaler v Comptroller-General of Patents, Designs and Trade Marks. The Supreme Court confirmed in December 2023 that only people can be inventors, and the High Court reinforced this in September 2025 addressed whether an artificial intelligence system could itself be named as an inventor on a patent application.

The Parties and Background

The claimant, Dr. Stephen Thaler, sought to patent inventions purportedly conceived by DABUS, his patented artificial intelligence system. Dr. Thaler claimed that DABUS had conceived inventions independently and autonomously, without human intervention. Among its purported inventions were a novel food container and a new type of light beacon designed to attract attention in emergencies.

Dr. Thaler submitted patent applications in the United States, European Union, and United Kingdom, listing DABUS as the sole inventor. He contended that the requirement to list a human inventor should either be expanded to include artificial intelligence systems or waived entirely. The respondent was the Comptroller-General of Patents, Designs and Trade Marks, representing the United Kingdom Intellectual Property Office.

Historical Context: DABUS and Machine Learning

Historically, machine-learning systems have been regarded as tools trained to perform specific tasks. Any invention arising from those tasks would not usually be considered an invention of the machine itself, as the outcome is predetermined by its programming and training data. DABUS challenged this understanding by purportedly generating solutions its creator did not directly program it to produce.

International Responses

In December 2021, the European Patent Office dismissed Dr. Thaler's application. In April 2023, the United States Supreme Court declined to hear his appeal, leaving in place earlier rulings that inventions must list a human inventor. Both the EPO and USPTO maintained that naming an artificial intelligence system as an inventor would "contradict the plain reading" of statutes referring to "persons" and "individuals."

UK Supreme Court Ruling in Thaler

Dr. Thaler's United Kingdom patent application, first submitted in 2018, did not name a human inventor. After years of legal proceedings through multiple appellate levels, the United Kingdom Supreme Court delivered its ruling in December 2023. The Court ruled unanimously that an inventor must be a natural person under the Patents Act 1977. The Court concluded that the term "inventor" refers exclusively to humans and cannot include machines or artificial intelligence systems. The judgment emphasized that any change to this statutory interpretation would require legislative reform by Parliament.

Subsequent UKIPO Guidance (2025)

Following the Supreme Court's judgment in Thaler, the United Kingdom Intellectual Property Office issued updated guidance in 2025 clarifying its position. The UKIPO reaffirmed that artificial intelligence systems cannot be listed as inventors and that patent applicants must identify a human inventor who "makes a material contribution" to the inventive concept. The High Court reinforced this position in September 2025, making clear that patent applications naming only an artificial intelligence system as inventor can be treated as withdrawn.

Practical Implications of the Thaler Decision

The Supreme Court's determination that only natural persons may be named as inventors has significant practical consequences. Individuals or businesses using artificial intelligence tools to develop inventions may still list themselves or their employees as the inventor, even if the human contribution was limited to setting up, training, or operating the artificial intelligence system.

Potential Challenges Avoided

The Court's decision avoided significant practical and legal challenges that would have arisen from recognizing artificial intelligence systems as inventors. Such recognition would have raised complex questions for litigation, including how a machine could be cross-examined, whether it could possess legal rights, and how ownership of inventions would be determined. The ruling maintains consistency with the established approach in Europe and the United States, avoiding fragmentation in international patent practice.

The Developing State of AI Patent Law

The two Supreme Court decisions together illustrate the developing nature of artificial intelligence patent law in the United Kingdom. The law now distinguishes clearly between two separate questions: who may be an inventor (only natural persons, per Thaler), and what types of AI-implemented inventions are patentable (significantly expanded scope per Emotional Perception AI).

This dual framework reflects a policy balance: encouraging innovation in artificial intelligence technologies by ensuring such inventions receive patent protection, while maintaining the traditional link between human creativity and inventorship. The framework acknowledges that artificial intelligence systems are increasingly capable of generating creative and technical solutions, yet preserves the requirement of human contribution to the inventive process.

Legislative and Policy Considerations

Both judgments emphasize that the current legal framework is established by statute, specifically the Patents Act 1977, and that any fundamental changes would require legislative intervention by Parliament. This places responsibility upon legislators to consider whether existing law adequately addresses the challenges posed by increasingly sophisticated artificial intelligence systems.

As artificial intelligence systems become more capable of autonomous problem-solving and creative output, legislators face the challenge of incentivizing human innovation while acknowledging the growing role of artificial intelligence in the inventive process. The Supreme Court's decisions provide clarity under current law but do not foreclose the possibility of future legislative reform.

Harmonization with International Practice

The United Kingdom's approach now aligns with major patent jurisdictions internationally. The requirement that inventors be natural persons is consistent with decisions in the United States and European Patent Office. Similarly, the adoption of EPO practice regarding patentability of AI-implemented inventions in Emotional Perception AI reduces divergence within the European patent system.

Impact on Patent Drafting and Prosecution

Patent practitioners must now navigate this dual framework when advising clients developing artificial intelligence technologies. Applications must identify human inventors who have made material contributions to the inventive concept, even where artificial intelligence tools played a significant role in generating solutions. At the same time, practitioners should emphasize technical implementation details, hardware involvement, and technical effects to ensure AI-implemented inventions satisfy the expanded eligibility criteria established in Emotional Perception AI.

The requirement to identify human inventors necessitates careful documentation of the development process, particularly in collaborative environments where multiple individuals and artificial intelligence systems contribute to inventive output. Organizations developing artificial intelligence technologies should establish clear policies regarding inventorship determination and maintain records sufficient to demonstrate human contributions.

Implications for Broader Intellectual Property Protection

While patent law now provides clearer guidance, questions remain regarding other forms of intellectual property protection for AI-generated outputs. Copyright law, trade secret protection, and design rights each present distinct questions regarding the role of artificial intelligence in creating protectable subject matter.

The Supreme Court's reasoning in Thaler, emphasizing that statutory language refers to natural persons, may influence the interpretation of inventorship and authorship requirements in these related contexts.

Continuing Evolution of AI Capabilities

The legal framework established by these decisions will be tested as artificial intelligence capabilities continue to advance. Current systems that require human setup, training, and operation fit comfortably within the framework: humans remain inventors, and AI-implemented systems are patentable. However, future systems with greater autonomy may renew questions about the adequacy of existing legal categories.

The Supreme Court's emphasis on the need for legislative reform to alter the inventorship requirement suggests that courts will not expand the definition of "inventor" through judicial interpretation, regardless of technological developments. This places responsibility upon Parliament to monitor technological progress and consider whether statutory amendments are necessary to maintain an effective patent system.

Comparative International Developments

While the United Kingdom, United States, and European Patent Office have reached similar conclusions regarding AI inventorship, some jurisdictions have adopted different approaches. Monitoring international developments will be important as businesses seek patent protection across multiple territories. Divergent approaches to AI-related patent issues could complicate international patent prosecution strategies and licensing arrangements.

Economic and Innovation Policy Considerations

The Supreme Court's decisions reflect underlying policy judgments about innovation incentives. By expanding patentability of AI-implemented inventions while maintaining human inventorship requirements, the framework seeks to encourage development and deployment of artificial intelligence technologies while preserving the patent system's focus on rewarding human ingenuity.

The economic significance of these decisions extends beyond individual patent applications. Clear legal rules regarding patentability and inventorship reduce uncertainty for investors, researchers, and commercial entities operating in the artificial intelligence sector. This clarity may encourage increased investment in United Kingdom-based artificial intelligence research and development.

Broader Implications for Technology Law

These decisions form part of a broader judicial engagement with artificial intelligence and emerging technologies. As artificial intelligence systems are deployed across sectors including healthcare, finance, transportation, and manufacturing, courts and regulators must address questions regarding liability, accountability, transparency, and fairness. The Supreme Court's approach in these patent cases, providing clarity under existing statutory frameworks while acknowledging the potential need for legislative reform, may inform judicial approaches to artificial intelligence in other legal contexts.

Observations on Technical Character and Hardware Implementation

The Emotional Perception AI judgment's focus on technical character and hardware implementation reflects a pragmatic approach to emerging technologies. By recognizing that artificial neural networks necessarily operate through physical computing infrastructure, the Court avoided formalistic distinctions based on implementation methods. This approach should accommodate future technological developments, including specialized artificial intelligence hardware, quantum computing implementations, or novel neuromorphic architectures.

The Court's technologically neutral characterization of artificial neural networks as computer programs, combined with its holding that hardware implementation confers technical character, creates a framework adaptable to evolving technology. This flexibility should reduce the need for frequent legislative or judicial intervention as artificial intelligence technologies continue to develop.

Conclusion

The United Kingdom Supreme Court's decisions in Thaler v Comptroller-General of Patents, Designs and Trade Marks (December 2023) and Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks (February 2026) establish a comprehensive framework for artificial intelligence and patent law in the United Kingdom.

The Thaler decision definitively establishes that only natural persons may be named as inventors on patent applications. This requirement reflects the statutory language of the Patents Act 1977 and can only be altered through legislative reform. The subsequent 2025 UKIPO guidance and September 2025 High Court decision reinforce this principle, clarifying that patent applications listing only artificial intelligence systems as inventors may be treated as withdrawn.

The Emotional Perception AI decision significantly expands the scope of patentable subject matter for AI-implemented inventions. By abandoning the Aerotel test and adopting European Patent Office practice, the Supreme Court lowered the threshold for technical character while maintaining substantive requirements of novelty, inventive step, and industrial applicability. This creates a more favorable environment for protecting artificial intelligence innovations in the United Kingdom.

Together, these decisions demonstrate the developing nature of artificial intelligence patent law. The framework distinguishes between inventorship (requiring human inventors) and patentability (permitting AI-implemented inventions with technical character). This dual approach balances innovation incentives with legal tradition and international harmonization.

Future Developments and Legislative Considerations

Both judgments acknowledge that the current legal framework is established by statute and that fundamental changes require parliamentary action. As artificial intelligence capabilities continue to advance, legislators will face questions about whether existing law adequately incentivizes innovation and appropriately allocates rights in AI-generated inventions.

The Supreme Court's decisions provide clarity and predictability under current law while explicitly recognizing the potential need for future legislative reform. This approach respects the separation of powers between judiciary and legislature while providing guidance to innovators, investors, and practitioners operating under existing legal frameworks.