The integration of artificial intelligence into legal workflows offers the promise of revolutionizing evidence gathering in complex litigation. For trademark practitioners, this potential is significant, particularly where visual evidence - such as product placements or retail signage - is critical to proving infringement. However, a recent ruling by the UK High Court illustrates that technological capability does not automatically equate to judicial approval. The application of AI tools in trademark law is governed not merely by what technology can achieve, but by whether its use is proportional to the stakes of the dispute.
In B&Q Limited v. Wren Kitchens Limited, Deputy High Court Judge David Stone declined an order that would have allowed one party to spend approximately £100,000 deploying an AI image-search tool across a competitor’s retail data. The decision underscores a fundamental principle in modern litigation: the cost of discovery must remain reasonable relative to the likely value of the evidence obtained.
The Context of Visual Trademark Infringement
The underlying dispute between Wren Kitchens and B&Q centers on allegations of comparative advertising. B&Q sued Wren for trademark infringement, arguing that promotional materials claiming Wren’s kitchens cost less than B&Q’s were misleading. Conversely, Wren counterclaimed that B&Q had infringed its trademarks by using comparison tables from a third-party website in its own marketing efforts.
A key challenge in such cases is establishing the scope and context of a competitor’s alleged trademark use. While traditional document disclosure relies heavily on keyword searches within text documents, much of modern retail advertising exists in visual formats: photos of store shelves, digital banners, and in-store signage. These elements often contain trademarks embedded in images rather than searchable text fields, rendering them invisible to standard e-discovery protocols.
Wren sought to bridge this gap by requesting an expansion of the disclosure order. They proposed using a proprietary AI tool capable of optical character recognition (OCR) and image analysis to scan over 600,000 documents from B&Q’s retail operations. Wren argued that this technology was essential to uncover instances where their trademarks appeared in images, which text-based searches would inevitably miss.
The Proportionality Doctrine in Action
Despite the sophistication of the proposed AI tool and a significant drop in its estimated cost - from £400,000 to £100,000 due to technological advancements - the judge rejected the request. The ruling hinged on the legal doctrine of proportionality, a cornerstone of civil procedure in many jurisdictions, including the UK.
The court assessed whether the £100,000 expenditure was justified by the potential evidential benefit. Several factors weighed against Wren:
- Diminishing Returns: The court noted that previous searches had already yielded relevant images, suggesting that further AI-driven discovery would likely produce marginal gains.
- Temporal Relevance: The proposed search period predated the core allegations by roughly twelve months, raising questions about its direct relevance to the pleaded issues.
- Existing Disclosure Burdens: The parties had already incurred nearly £1.4 million in disclosure costs, with B&Q having processed hundreds of thousands of documents. Adding a six-figure cost for speculative discovery was deemed unreasonable.
Judge Stone emphasized that while AI tools can enhance the thoroughness of evidence gathering, their use is not automatic. The court must balance the necessity of the search against its cost and the likelihood of finding material that changes the outcome of the case.
Implications for Trademark Practitioners
This decision carries significant weight for businesses and legal teams navigating trademark disputes in an increasingly visual marketplace. It signals that courts are becoming more skeptical of high-tech, high-cost discovery motions unless they are tightly scoped and demonstrably necessary.
For trademark owners, the implication is clear: relying on AI to cast a wider net does not guarantee access to evidence. Litigation strategy must focus on identifying precise gaps in existing disclosure rather than seeking broad, technology-enabled sweeps. Parties must be prepared to justify the cost of AI interventions with concrete arguments regarding relevance and necessity, rather than simply pointing to the tool’s capabilities.
Conversely, for businesses facing infringement allegations, this ruling offers a layer of protection against excessive and speculative discovery practices. It reinforces the right to limit disclosure obligations when they become disproportionate, ensuring that the cost of defending intellectual property rights does not eclipse the value of the dispute itself.
Balancing Innovation with Legal Pragmatism
As AI technology continues to evolve and costs decrease, the threshold for its acceptance in court may shift. However, the core legal principles governing discovery remain stable. Courts will continue to prioritize efficiency and fairness over technological novelty.
For trademark practitioners, the lesson is not to avoid AI but to use it strategically. The most effective applications of AI in litigation are those that address specific, identified deficiencies in evidence rather than those aimed at speculative fishing expeditions. By grounding technology-driven discovery requests in rigorous legal justification, lawyers can ensure that innovation serves justice without compromising the economic realities of commercial litigation.
The ongoing trial between Wren and B&Q will ultimately determine the substantive claims of infringement, but the procedural precedent set here will likely influence how AI is utilized in future trademark disputes. The message is unequivocal: technology must be wielded with discretion, guided by the enduring legal imperative of proportionality.