UK Design Law Consultation Spotlight Series: Computer-Generated Designs

The UK Intellectual Property Office (UKIPO) is consulting on significant reforms to the legal framework for designs. In this fourth article in our Spotlight series, we look at Section D of the consultation, addressing how the law should treat computer-generated designs (CGDs) made without a human author, particularly in light of rapid advances in generative AI and the UK’s broader policy direction on AI and IP. The consultation is open until 27 November 2025.

Computer-Generated Designs — Background and Rationale

An increasingly salient issue in designs is whether and how UK design law should protect designs that are generated by a computer in circumstances where no human is the “author”. UK registered design law and UK unregistered design law currently include provisions that can protect such designs by deeming the “author” to be the person who made the arrangements necessary for their creation. Supplementary unregistered design (SUD) law contains no equivalent provision. These CGD provisions mirror the long-standing UK approach in copyright to computer-generated works. This fragmented position causes uncertainty, invites forum-specific outcomes, and is increasingly out of step with wider European developments that refuse or limit protection for purely computer-generated outputs.

AI systems can generate design outputs at industrial scale and speed. This heightens the risk that mass, low-quality filings could erode novelty and individual character, elevate clearance burdens, and crowd out human designers. There is also a risk of undermining the value of human creativity in design, if registered designs become a ‘first to file’ commodity regardless of any intellectual investment. The consultation proposes aligning UK law with the EU’s human-centred approach, which has already excluded designs without a human author from protection. Against this backdrop, the consultation’s preferred approach is to remove CGD protection for designs without a human author.

Proposals for Reform

The consultation puts forward four main options:

Maintain existing CGD protection (and consider extending an equivalent provision to SUD for consistency). This option is not preferred given the risks to human creativity, the lack of evidence of use or need, and divergence from EU policy.

Reform rather than remove: Adjust the framework to address conflicts, particularly the originality requirement for UK unregistered designs (for example, redefining or disapplying originality for CGDs). This would add legal complexity and could blur the human-centric boundary that policy is moving towards.

The UKIPO’s preferred option: Remove the CGD provisions for designs without a human author: Repeal the deeming provisions from registered and UK unregistered design law and maintain the current position for SUD. This would align the UK with the EU’s human-centred approach, reduce the risk of mass CGD filings degrading novelty and individual character, and resolve the originality conflict in UK unregistered design.

Supplementary option — Collect AI-usage data at filing: Require or invite applicants to disclose whether AI was used in creating the design, and potentially to specify the tools, the degree of AI involvement, or the nature of the human contribution.

The fourth option can be implemented alongside any of the other options listed above. This could be as simple as asking applicants to make a declaration whether they had used AI in the generation of the design and the name of the AI tool used in the creation of a design.

Our Views

We agree with the UKIPO’s preference, that the existing computer-generated designs provisions should be removed. The current law actually creates unpredictability and risk for (human) designers, since generative AI allows the production of matter at a scale far beyond what was envisaged when the current provisions were drafted in 1988. Ultimately, removing protection for purely computer-generated designs will hopefully preserve the integrity of UK design law while still enabling the responsible use of AI as a creative aid. This position aligns with the government’s proposal to delete the equivalent provision in its copyright/AI consultation, although no firm decision on that has been reached.

It makes little sense to have different rules for the different design rights, and the current inconsistency with SUD is more of a historical accident (owing to Brexit) than the result of a careful legislative scheme. Moreover, a recurrent theme throughout these reforms is the desirability of aligning with the EU, where equivalent protection has been explicitly denied. A unified approach will hopefully facilitate cross-border business operations and will help courts and registries apply a coherent standard when assessing novelty and individual character in AI-generated works.

However, removing the provision would not put an end to the matter of CGD. It will remain unclear as to what extent of human involvement is required for a design to be considered “authored” by a natural person. For example, if using generative AI, is writing a prompt sufficient – or does it depend on the scope and complexity of the prompt? Would selection from a list of AI-generated designs count as human involvement? Or would human involvement have to occur in the nature of active intervention or creative input into the design, and if so, to what extent? The ultimate arbiters of these questions will likely be Parliament or the courts, but the design community will naturally look to the UKIPO for high-quality practical guidance.

As to collecting supplementary information, the question is more finely balanced. Collecting information about AI use could enhance transparency, clarify questions of authorship and ownership, and provide a valuable evidence base for future policymaking. However, if the CGD provisions were to be removed, this extra information disclosure is unlikely to assist examination and might add friction to a system prized for speed and simplicity. In any event, crude tick-box disclosures or percentage allocations are unlikely to resolve substantive questions that will ultimately be litigated.  More generally, the growing focus on the question of “bad faith” in UK design applications also begs the question of whether applicants could be trusted to make truthful and honest declarations when the grant of protection depends on it. Without a meaningful sanction for false declarations (over and above the risk of invalidity, which is always present), an application process which relies solely on self-disclosure would be open to abuse. Further thought would also be needed on practical ways to deter dishonest filings by imposing more robust evidential requirements to document the process of creation, and facilitating post-grant invalidity for designs suspected of lacking any human authorship.

Article co-authored by Ana-Maria Curavale, Trainee Solicitor