Intellectual Property Law Dissertation Topics

You have picked intellectual property law as your dissertation area, but IP is really four different bodies of law living under one label: copyright, patents, trade marks and trade secrets. Which one is your research question actually about, and can you defend that choice in your proposal? Is the topic you are drawn to a settled area of law you can analyse doctrinally, or a live reform question where the position could shift before you submit? Do you have realistic access to case reports, IPO decisions or licensing data, or are you assuming sources exist that you have not actually checked? These are the questions worth settling before you commit to a title.

This resource covers all four branches of UK intellectual property law, using the Copyright, Designs and Patents Act 1988, the Patents Act 1977, the Trade Marks Act 1994 and the Trade Secrets (Enforcement, etc.) Regulations 2018 as the core statutory framework, with EU and international law brought in only where it genuinely shapes the UK position. A large share of the most researchable current material sits at the intersection of IP law and artificial intelligence, so several topics below draw on that, but the list is not limited to it.

Position correct as of September 2026. IP and AI law is moving fast: the government’s Report on Copyright and Artificial Intelligence was only published in March 2026, and the Getty Images v Stability AI judgment is under appeal. Check the Intellectual Property Office and the relevant court’s website for the current status before you finalise a proposal built on any of these developments.

Copyright law and artificial intelligence

1. Does secondary infringement under sections 22 and 23 of the CDPA 1988 meaningfully restrain AI model distribution after Getty Images v Stability AI [2025] EWHC 2863 (Ch)?
Suitable for Master’s or PhD level. In November 2025 the High Court held that Stable Diffusion did not constitute an “infringing copy” because the model’s weights do not store or reproduce Getty’s images, even though Mrs Justice Joanna Smith accepted that “article” under the CDPA can extend to intangible, electronically stored objects. A dissertation could examine what this leaves open for AI developers who do train models on UK soil, since Getty’s primary infringement and database right claims were withdrawn for lack of UK jurisdictional nexus rather than decided on the merits. Evidence route: the full judgment, the Court of Appeal’s eventual ruling on the granted permission to appeal, and academic commentary tracking the parallel US cases (Bartz v Anthropic, Authors Guild v OpenAI). Common pitfall: treating the case as settling generative AI copyright law generally, when the judge herself described the outcome as fact-specific.

2. What should replace the rejected opt-out text and data mining exception, following the UK government’s March 2026 Report on Copyright and Artificial Intelligence?
Suitable for Master’s or PhD level. The government’s original preferred option, a broad TDM exception with a rights-reservation opt-out modelled on the EU approach, was abandoned after most consultation respondents rejected it. The report instead points to a possible narrower, use-specific exception as the most popular alternative. A strong dissertation would evaluate the four original options against the reporting duty in sections 135 and 136 of the Data (Use and Access) Act 2025, and assess whether a targeted exception can satisfy both creative-industry licensing interests and AI developers without repeating the opt-out’s practical difficulties. Framework: compare against the EU’s Article 4 Digital Single Market Directive exception, which the UK proposal borrowed from.

3. Can section 9(3) CDPA’s computer-generated works provision coherently attribute authorship to generative AI outputs?
Suitable for undergraduate or Master’s level, explained in plainer terms at undergraduate level. Section 9(3) attributes authorship of a computer-generated work to “the person by whom the arrangements necessary for the creation of the work are undertaken,” a rule drafted in 1988 for very different technology. A dissertation could test whether this wording sensibly identifies an author when a generative model, not a human arranger, does most of the creative work, and compare the UK’s unusual retention of this provision against the EU and US, neither of which has an equivalent. Sample research question: does section 9(3) protect AI-assisted outputs in a way that is consistent with the originality threshold applied to human authorship elsewhere in the CDPA?

4. Should UK copyright law adopt AI training-data transparency duties comparable to Article 53 of the EU AI Act?
Suitable for Master’s level. The UK’s March 2026 report favoured transparency measures even while shelving the broader exception question. A dissertation could assess what a workable UK transparency obligation would actually require developers to disclose, whether it should apply extraterritorially given that most large models are trained outside the UK, and what enforcement mechanism would give it teeth. Feasibility note: this is a policy-design question best tackled through comparative doctrinal analysis and consultation-response review rather than empirical data collection, since no UK transparency regime yet exists to study in practice.

5. How feasible is a collective licensing or remuneration scheme for creators whose work trains commercial AI models?
Suitable for PhD level. Rather than an exception-based approach, some stakeholders in the 2024 to 2025 consultation argued for extending collective licensing bodies to cover AI training use, similar to existing music and text licensing schemes. This topic could examine whether the infrastructure used by organisations such as the Copyright Licensing Agency could realistically be adapted for AI training at scale, drawing on transaction-cost economics and the empirical licensing evidence submitted during the consultation. A common pitfall is assuming collective licensing solves the problem without addressing how a foreign-trained model would ever be brought within a UK licensing scheme’s reach.

Patent law

6. Why did Thaler v Comptroller-General [2023] UKSC 49 close the door on AI systems as patent inventors, and what gap does this leave for autonomously generated inventions?
Suitable for undergraduate or Master’s level. The Supreme Court unanimously held that “inventor” under sections 7 and 13 of the Patents Act 1977 must be a natural person, so DABUS could not be named and Dr Thaler could not claim entitlement merely by owning the machine that generated the inventions. A dissertation could analyse whether this leaves genuinely autonomous inventions unpatentable in the UK altogether, forcing applicants to attribute inventorship to a human “arranger” in a way that may not reflect what actually happened, and compare this to the different outcomes reached in Germany, Australia and the United States on the same DABUS applications.

7. How should section 39 Patents Act 1977 employee-ownership rules apply when staff use generative AI tools to develop a patentable invention?
Suitable for Master’s level. Section 39 gives employers ownership of inventions made in the normal course of an employee’s duties, but that framework assumes the employee is the inventor doing the inventive work. A dissertation could examine how far an employee who prompts and curates an AI system’s output, rather than conceiving the invention unaided, still satisfies the inventive-contribution threshold the courts require, and what this means for employment contracts and R&D governance in AI-heavy industries. Evidence route: IPO hearing officer decisions on inventorship disputes, employment tribunal case law on inventor remuneration under section 40, and interviews with in-house patent counsel if ethics approval and access allow.

8. Does the UK’s continued non-participation in the Unified Patent Court disadvantage UK-based innovators seeking European-wide protection?
Suitable for undergraduate or Master’s level. The UK confirmed in 2020 that it would not join the Unitary Patent or Unified Patent Court systems, both of which came into operation for participating EU states from 2023. A dissertation could assess whether UK businesses now face higher costs and greater complexity securing pan-European protection through the traditional national-validation route, using patent-filing and litigation-cost data where it is publicly available, and weigh this against the UK’s continued access to the European Patent Office for individual national patents.

9. How consistently do the UK IPO and the EPO apply the computer program exclusion under section 1(2) Patents Act 1977 to AI-related inventions?
Suitable for Master’s or PhD level, requiring denser technical argument. Software “as such” is excluded from patentability, but AI-related inventions routinely secure patents where they demonstrate a technical contribution beyond the program itself. A dissertation could build a small comparative sample of UK IPO and EPO decisions on AI-related applications, applying the established Aerotel/Macrossan and EPO problem-solution approaches, to test whether the two offices are converging or diverging in how they treat machine-learning inventions. Feasibility: published examination guidelines and decided cases are freely accessible, making this a realistic doctrinal comparison without primary data collection.

Trade mark law

10. Should the UK’s decision to keep the UK+ exhaustion of IP rights regime be revisited now that its 2025 review has closed the question?
Suitable for undergraduate or Master’s level. Since 2021 the UK has unilaterally recognised exhaustion of rights for goods first placed on the EEA market, without the EU reciprocating for UK-origin goods, and the government confirmed in 2025 that this asymmetric “UK+” regime will continue. A dissertation could evaluate the practical effect on UK rights holders’ ability to stop parallel exports into the EEA, using the government’s own consultation responses and sector-specific evidence (pharmaceuticals, luxury goods, publishing) where publicly available, and assess whether the stated reason for retaining the status quo, a lack of reliable economic data, still holds.

11. What does Getty’s narrow trade mark win under sections 10(1) and 10(2) Trade Marks Act 1994 reveal about liability for AI-generated brand replication?
Suitable for Master’s level. While Getty lost its copyright claims against Stability AI, the court found limited trade mark infringement where early versions of Stable Diffusion reproduced Getty and iStock watermarks in generated images. A dissertation could examine what this narrow finding suggests about the boundary between an AI developer’s liability for its model’s design choices and a user’s liability for the specific prompts they enter, and whether trade mark law is better suited than copyright to catching certain categories of AI-generated brand harm.

12. How well does UK trade mark classification under the Nice Agreement accommodate virtual goods and NFT-linked brand use?
Suitable for undergraduate level, explained accessibly. Brand owners have increasingly sought registrations covering virtual goods and non-fungible tokens, and UK IPO practice has had to interpret existing Nice classes for products that have no physical form. A dissertation could review a sample of UK trade mark applications and oppositions involving virtual goods specifications, assess whether current classification guidance gives applicants and examiners enough certainty, and compare the UK IPO’s approach to EUIPO guidance on the same question.

Trade secrets and confidential information

13. Where does the boundary sit between breach of confidence and the Trade Secrets (Enforcement, etc.) Regulations 2018 for UK businesses?
Suitable for undergraduate or Master’s level. The 2018 Regulations implemented the EU Trade Secrets Directive alongside, rather than replacing, the common law action for breach of confidence, leaving businesses with two overlapping routes to protect the same information. A dissertation could map where the statutory definition of a “trade secret” (information that is secret, has commercial value because it is secret, and has been subject to reasonable steps to keep it secret) diverges from the more flexible equitable duty of confidence, and identify which route gives a claimant a practical advantage in a typical departing-employee dispute.

14. How effectively do post-termination restrictive covenants protect trade secrets when employees move to AI-driven competitors?
Suitable for Master’s level. Non-compete and non-solicitation clauses are the practical front line for protecting confidential information when key staff leave, but courts will only enforce covenants that go no further than necessary to protect a legitimate business interest. A dissertation could examine how this restraint-of-trade doctrine is being tested in fast-moving AI and data-driven sectors, where the “legitimate interest” (proprietary model architecture, training data, algorithmic know-how) is harder to define than in traditional client-relationship cases, drawing on recent High Court covenant enforcement decisions.

15. What trade secret risks does generative AI create through model inversion and prompt-based data extraction?
Suitable for Master’s or PhD level. Security researchers have demonstrated that some AI systems can be prompted or probed in ways that reconstruct fragments of their training data or reveal proprietary system prompts, raising a genuinely new category of trade secret exposure that existing confidentiality frameworks were not built to address. A dissertation could assess whether the “reasonable steps to maintain secrecy” requirement in the 2018 Regulations can sensibly be applied to information embedded inside a trained model, and what technical and contractual safeguards a business would need to show it took. This is a genuinely open research area, so keep claims about legal certainty cautious and clearly framed as scope for future clarification rather than settled law.

IP licensing and commercialisation

16. How is the AI training-data licensing market reshaping traditional IP licensing practice in the UK?
Suitable for Master’s level. As litigation risk around unlicensed AI training grows, publishers, image libraries and data holders have increasingly moved toward direct licensing deals with AI developers rather than relying on litigation or a future statutory exception. A dissertation could analyse how these emerging agreements differ in structure from traditional copyright licensing (pricing models, exclusivity, ongoing royalty versus lump sum), using publicly reported deals and industry commentary as the primary evidence base, since full contract terms are rarely public.

17. What IP ownership and licensing challenges arise in university spin-out and technology transfer agreements?
Suitable for undergraduate or Master’s level. Publicly funded research generates patentable and copyrightable outputs whose ownership is governed by university IP policies, funder terms and, for collaborative projects, industry sponsorship agreements that can conflict with each other. A dissertation could compare a sample of UK university technology transfer office policies, assess how they allocate rights between academic inventors, the institution and any commercial partner, and identify where ambiguity commonly causes disputes at the point of spin-out.

Comparative and cross-cutting

18. How do the UK, EU and US diverge on text and data mining exceptions after Bartz v Anthropic and the UK’s stalled reform?
Suitable for PhD level. The UK has no general TDM exception for commercial use and, as of the March 2026 report, no settled plan to introduce one; the EU has a conditional exception with an opt-out; the US relies on the uncertain and actively litigated fair use doctrine, with mixed outcomes emerging from cases against Anthropic, OpenAI and others. A dissertation could build a structured comparative framework across the three jurisdictions, assessing which approach currently offers AI developers the most legal certainty and which offers rights holders the strongest practical protection, with the explicit caveat that all three positions may move again before submission.

Choosing a doctrinal approach that fits your topic

Most IP law dissertations are doctrinal at their core: close statutory interpretation combined with case analysis. Where a topic asks how courts or the IPO have actually applied a provision, a case-law method built around a defined sample of decisions is usually stronger than a broad literature review, because it lets you show independent legal analysis rather than summarising secondary commentary. Where a topic is genuinely comparative, such as the TDM divergence or the computer-generated works question, be explicit from the outset about which jurisdictions you are comparing and why, rather than adding foreign law as decoration.

A common mistake in IP dissertations is choosing a topic because it is trending in the legal press without checking whether enough primary material exists to sustain 10,000 to 15,000 words of original analysis. A single High Court judgment, however high-profile, rarely supports a whole dissertation on its own; you need either a body of related case law, a live statutory reform process with documented consultation responses, or a comparative structure that gives you enough material across jurisdictions. Before committing, check that the case reports, IPO decisions or government documents you plan to rely on are actually accessible, and that any pending appeal or ongoing consultation will not make your central argument obsolete before your submission date.

If you are narrowing from one of the broader topics above, the free dissertation topic finder and research question generator can help turn a category like “copyright and AI” into a single, focused, researchable question. For structuring a formal proposal once you have settled on a direction, our subject dissertation help for law covers planning and feedback at every stage.

If your interest sits closer to digital regulation than IP specifically, the artificial intelligence dissertation topics resource covers the wider AI governance and liability landscape. Students weighing a commercial law angle on licensing and technology contracts may find the commercial law dissertation topics resource useful, and the exhaustion and parallel-import topics above connect naturally to competition law dissertation topics. For the wider law subject overview, see law dissertation topics.