AI Statement
This is my artificial intelligence philosophy statement — it’s intended to give clients, colleagues and members of the public a straightforward picture of how I use AI in my work — and, just as importantly, how I don’t use AI.
AI has been part of my life since 2018, when I became one of the first lawyers in the world to specialise in deepfakes. I advise on it, speak about it, experiment with it and, yes, occasionally have a mini existential crisis about it. That’s precisely why it matters to me to have clear principles and boundaries around how I use it.
This isn’t marketing fluff about whatever happens to be considered trendy, innovative or ethical at any given moment. Nor is it a generic promise to use AI “responsibly” without saying what that actually means. Instead, it’s about being genuinely transparent.
The lawyer’s favourite answer is “it depends” — because context really does matter. So below, I’ve tried to set out as clearly as possible where AI fits into my work, and where it doesn’t.
Legal work for clients
Provided always that no confidential information is uploaded, I may use AI to:
✓ Assist with non-sensitive research, subject to my own independent verification
✓ Stress-test my thoughts and ideas around a particular problem or document
✓ Consider certain commercial issues relevant to my client, like market insights
✓ Organise my workload, priorities and to-do lists
★ In special circumstances, I may intentionally stress-test risks and vulnerabities associated with particular AI tools, for example to see whether a model’s guardrails will prevent it from generating unlawful content.
I never use AI to:
✗ Draft specific wording used in contracts
✗ Write emails, not even with Microsoft Copilot or Gmail Gemini auto-complete features
✗ Organise or otherwise access to client emails, communications or files
✗ Analyse privileged, confidential or sensitive client information
✗ Make decisions or replace my professional judgement
Content & Branding
I may use AI to:
✓ Develop outlines, structures and ideas for presentations, social media and marketing
✓ Improve my online engagement and website content
✓ Brainstorm examples, titles or ways of explaining complex topics
✓ Generate images and other audiovisual material for presentations and marketing
I never use AI for:
✗ Publishing AI-generated material without my own substantive review and input
✗ Generating synthetic versions of identifiable people’s voices, likenesses or performances without appropriate consent
The AI tools I use
When I use AI, I choose tools that:
✓ Are provided by established companies based in the UK, European Union or United States
✓ Have terms and conditions, privacy terms and data-use policies that I have reviewed
✓ Allow me to use appropriate privacy, security and data-control settings