| AI & ACCOUNTABILITY
ENGLAND
KEY TAKEAWAYS
Mind over
In the absence of dedicated AI legislation in the UK, businesses must rely on existing IP, copyright, and data protection laws. Responsible adoption involves implementing internal governance, auditing training data, and securing rights through carefully drafted agreements.
Q3 Are you seeing any trends in AI-driven disputes or liability concerns? How can firms assist clients in addressing potential AI-related litigation or regulatory scrutiny? AI poses many questions regarding liability. You have the problem of determining accountability; who is responsible if the AI systems causes harm? Is it the developer, the system or the users? A further question currently under discussion in Sweden is liability and how existing Swedish law and product liability law apply to AI. Can it be treated like a defective product under liability laws? Finally there is the question of ethical and legal compliance and how traditional legal concepts like negligence or fault apply to AI systems. How prepared are insurance companies at managing risks associated with AI? Sweden is actively working to address the challenges of AI liability through legal, regulatory, and policy measures. Sweden has had a national AI strategy since 2018: “National approach for artificial intelligence”. The strategy outlines principles for the responsible use of AI, including transparency, fairness and accountability. It also points out the general direction and focuses on the priorities; education, training, research, innovation framework and infrastructure. These principles are intended to guide the development and deployment of AI technologies and address potential liability issues. Further Swedish government agencies, such as the Swedish Data Protection Authority and the Swedish National Board of Trade, are actively involved in discussions about AI regulation and liability. They provide guidance and oversight to ensure that AI systems comply ethical standards. While specific AI liability laws (like the EU AI Act) are still under development, in Sweden existing laws are being interpreted and adapted to address AI-related issues. Sweden’s adherence to GDPR provides a strong for data and security which is an aspect of AI liability.
necessary arrangements for the creation of the work. This provision has yet to be clearly tested in court in the AI context, and ambiguity remains about who qualifies as the “arranger” – the developer, the user, or the AI trainer? There is also the issue of who is responsible in the instance the generated output infringes rights of a third party (whether patent, copyright or moral rights). The same principle applies here – the person responsible for the arrangements is generally held to be liable. WP Thompson can help mitigate these risks by (i) conducting due diligence on training data sources and advise on licensing requirements; and (ii) monitoring regulatory changes in AI and IP law in the UK and Europe as well as advising on how to structure AI development contracts to assign and retain rights lawfully.
machine
Navigating AI risks in a legal grey zone
AI systems often rely on large datasets that may include protected works. Businesses
risk infringement if material is used without proper licensing. Output-related issues are equally complex, with current UK law offering only limited clarity on ownership and liability for AI-generated works.
As litigation around AI-generated content grows, businesses must scrutinise indemnity clauses
and liability provisions in AI vendor agreements.
Lewis Mullholland Partner, WP Thompson
AI implementation carries significant IP risks at both the input and output stages of AI model development. As the old adage goes – you get out what you put in, and the same is true for AI generated works. The quality of the output very much depends on the quality of the input training data on which the AI model is run, with larger, better-quality datasets typically producing the best output results. The issue is that much of these large bodies of data are comprised of IP protected works. In the UK, the use of protected material without the proprietor’s permission may constitute copyright or database infringement, especially where the whole or a substantial part of a work is involved, and the limited safe harbour provided by the UK’s text and data mining exemption exist only for non- commercial uses, where lawful access to the work has already been granted. Similarly, ownership and licensing issues can arise when datasets are aggregated or transferred across borders, potentially violating local laws. Once satisfied the input complies with UK IP and data protection frameworks, users then need to ask – who actually owns the output data and AI-generated work? Generative AI tools present a new frontier of complexity regarding the ownership and infringement risks of AI-generated content. Under current UK law, the copyright of a computer-generated work (where there is no human author) belongs to the person who made the
As artificial intelligence (AI) continues to reshape business operations globally, UK enterprises find themselves at the crossroads of opportunity and legal complexity. Nowhere is this more evident than in the field of intellectual property (IP), where evolving use cases for AI challenge long-standing principles of ownership, infringement, and liability. Against this backdrop, IP firms play a critical role in helping businesses adopt AI responsibly, safeguard their assets, and navigate emerging regulatory and litigation landscapes. Q1 How can businesses in your jurisdiction adopt AI and automation responsibly, and what guidance are you offering to ensure regulatory compliance? The rapid pace of AI development is pressuring legal systems to adapt. At present, businesses must operate in an environment where many jurisdictions, including the UK, rely on existing IP and data protection frameworks, often
in the absence of AI-specific legislation. However, this does not diminish the need for proactive legal compliance. As a practical point, businesses governance policies that include ethical considerations and data use protocols; (ii) audit all data and content used in training any such AI systems, ensuring the legality of sourcing and compliance with copyright, confidentiality, and data protection regulations; and, (iii) secure proprietary rights to any output through well-drafted agreements with AI developers and data providers. The incoming AI legislation promised by the UK’s current looking to adopt AI responsibly should (i) establish internal AI government is expected to bring further clarity in 2025. Until then, responsible AI adoption must be built on strategic foresight and ongoing legal advice. Q2 What are the key risks of implementing AI, from data privacy to ethical concerns, and how can you help businesses in your jurisdiction navigate these complexities?
“As the old adage goes – you get out what you put in, and the same is true for AI generated works.”
ABOUT US...
wpt.co.uk
Lewis Mullholland is a Partner at WP Thompson, a Chartered (UK) and European Patent Attorney and a registered representative before the Unitary Patent Court (UPC), providing full intellectual property services in the UK and Europe. Lewis’ practice encompasses both contentious and non-contentious proceedings, with a particular focus in the mechanical, electrical and software technical fields. He has an in-depth knowledge of the patentability and enforcement issues surrounding digital media and telecommunication technologies, AI, software and business method inventions in the UK and Europe.
Established in 1873, WP Thompson is a specialist full IP prosecution and litigation firm with primary offices in London, the financial and legal centre of Europe, Liverpool, in the commercial and industrial heartland of the UK, and Cardiff, the capital of Wales and near the home of the UK Patent Office. With a strong direct client base in Europe and a wide range of expertise in our team, we offer the highest standard of advice and assistance to our clients. Our collaborative approach means that we work closely with our clients, getting to know their businesses and technologies so we can advise them from a position of knowledge and experience.
+44 151 236 6688 lmu@wpt.co.uk irglobal.com/advisor/lewis-mullholland
irglobal | 39
38 | irglobal
Made with FlippingBook - professional solution for displaying marketing and sales documents online