AI is in almost every client conversation ...

AI now comes up in almost every conversation I have with clients. It usually appears in discussions about cost cutting, efficiency and risk management.

The questions I am hearing most often are:

• How can we use AI to save time and money while keeping our legal liability under control?
• What happens if AI gets something wrong and who is legally responsible?
• Can we insure against AI failures?
• What happens if the AI bubble bursts and our supplier goes bust after we have started relying on their technology?

These are sensible questions and my answers tend to fall into three key areas.

1. The small print matters

AI use is now becoming routine but I think that this makes the AI contract terms even more important. I think that many organisations are adopting AI without fully understanding the terms attached to using AI systems.

Key issues include whether the provider uses your data for training, how long data is stored and where and whether or not the provider has excluded responsibility for incorrect outputs or limited its liability to a level that offers little real protection or rights of redress.

These points directly affect legal risk.

2. Insurance is still catching up

The insurance market is developing AI-related cover but this cover is far from complete. 

Some policies address cyber incidents, professional negligence or technology-related crime but there doesn’t seem to be a single insurance policy that covers all the ways that using AI can create legal liability.

Importantly, insurance will not cover regulatory fines in many cases, and it will not help if the AI simply does not perform as well as you expected or does not deliver commercially.

I think that insurance should not be seen as a safety net for weak procurement or poor governance.

3. You need a fallback plan

I am seeing that dependence upon a single AI provider is becoming more common. That creates legal and business risk if the provider changes terms, is acquired or simply goes out of business. A fallback plan is essential.

This might include clear exit rights, data portability, use of alternative AI tools or a manual process you can revert to if and when required. 

Legal AI Seminar 

If you would like to hear more about how AI is being used in practice, what it can do, the legal issues that are arising and the trends I am seeing across organisations, then please do attend my AI seminar: The Impact of AI on Commercial Contract Drafting: Opportunities, Risks and Best Practices

I will cover real examples, explore where AI can go wrong, the common mistakes and practical steps you can take now to manage AI-related legal risks.

 

Published on Mar 19, 2026 by Angela Spall