AI and the law- Contracts, liability, and navigating the EU AI Act

I recently advised a longstanding UK-based software development and consultancy client of mine, on the legal complexities of developing an artificial intelligence (AI) system for deployment in the European Union.

The work

My client has developed an interactive AI system intended to enhance the visitor experience to an Italian public space.

The AI-powered tool serves as a digital guide, using adaptive technologies to take prompts from and learn about individual visitors as they move through the space, offering a personalised and tailored experience.

The need to Understand the European landscape

While the UK currently does not have specific AI legislation, the EU is the first territory in the world to legislate deployment of AI, with the Artificial Intelligence Act, which came into force on 1 August 2024, and is currently being rolled out in phases.

Although based in the UK, and acting as a sub-contractor, my client’s production of an AI system for deployment in the EU market brings it within the scope of the AI Act.

While the UK is no longer part of the European Union, the AI Act provides an early look into key concepts when unpacking the legal hurdles of AI models. These include:

· Assessing the level of risk associated with the AI system (e.g. the use of personal data or biometric identifiers).

· Clarifying the client’s role within the chain – whether as a producer, deployer, distributor or importer – which directly impacts their legal obligations and potential liabilities.

Contractual clarity

It is important to:

· Determine liability events and allocating them between the parties

· Consider Data protection considerations under GDPR and the AI Act

· Consider Intellectual property rights, particularly around how the AI is trained and whether it might generate original content or inventions

· Monitoring and human oversight

Technology will always outpace regulation – and AI is evolving at a pace never seen before in technology adoption.

What we can do is apply core legal fundamentals and anticipate likely risks to help our clients make informed, strategic decisions now.

Underlying legal principles – product liability, data protection, contract law, intellectual property – are not new, how they apply in an AI context is.

It is important to understand the AI system itself, its function, and its operational context before offering advice.

You need to take a forensic look at what it’s doing, to whom, and who’s involved in building and deploying it. From there, you tailor the legal framework to match.