AI creative and the law, how we help clients sleep soundly

Polarising views on AI are the loudest. But creative and marketing teams are getting on with it, using the technology and working through the conversations with their agencies about how to protect IP and reputations. At Love & Logic, we consider ourselves an AI-enhanced agency – and these are the conversations we’re having with clients every day.

Conversations about how we protect visual, asset, strategic, customer and market data and ensure we’re not inadvertently crossing the legal line by infringing third-party rights in our design work. And the measures that ensure complete transparency on how content is created (we get onto labelling later).

This is not legal advice. We wanted to share insights on the conversations you need to be across and the good practice we follow to protect ourselves and our clients. We stick to what we’re good at. Brand strategy and creative. Always seek legal advice from lawyers who specialise in AI and protecting brand IP.

It’s not straightforward. There’s no single law in the UK that governs AI – instead it sits across a patchwork of IP, advertising and consumer protection law. Enforcement is still finding its feet, and if your business operates across borders, the rules will differ by jurisdiction. Here’s what you need to know in the UK.

AI creative work – how do we protect IP?

In the UK, for now, AI creative work is defined as one of two things:

1. AI-assisted works (you have creative control, make creative choices and use your skill and judgement in producing work. You aren’t just using a basic prompt in Midjourney)
2. Computer-generated works (no human input beyond a basic prompt)

1. AI-assisted works – AI is part of our workflow, not the whole of it. Some elements of a project may be AI-generated, but we’re in complete creative control throughout – prompts, edits, curation, the direction of a project or campaign. That human creative control is what matters legally. It’s what qualifies the work as AI-assisted, gives it full copyright protection, and ownership transfers to you as part of our agreement.

Our value isn’t just in making things. It’s having the skill and judgement to know what to make in the first place. The insight, cultural fluency and taste buds to solve problems and create relevancy. AI can pump out 1,000 design options. It can’t tell you which one matters. Without human creativity, AI is just a statistical pattern engine that runs on training data. Work without any human input or oversight has a short shelf life – in terms of legal protections and consumer trust.

2. Computer-generated works – the UK is an outlier in protecting computer-generated works. The author and therefore rights owner is the person ‘who undertakes the arrangements necessary for the creation of the work’. If you wrote the prompt or set up the AI system you made the arrangements necessary. Copyright lasts for 50 years. But the UK Government said in March this year that it is leaning towards removing this protection to ‘incentivise and protect human creativity’. Be prepared.

Good practice – a few tips

Maintain human oversight – establish a clear process for human oversight and sign-off. Run your usual IP and trademark due diligence on AI creative. Use the same quality control checks that you would on your other design work, content and copy.

Documentation – document your prompts, iterations and human edits.

Platform Ts&Cs – read the platform Ts&Cs and check what warranties and indemnities are in place.

Where things get messy

• Working across borders – as mentioned earlier, you need to be across the legal frameworks in the jurisdictions where you work.
• Platform Ts&Cs – do you own the copyright or the platform? Check the Ts&Cs – read more here.

Where do things stand on labelling content and transparency?

There aren’t any regulations in the UK on labelling content that is generated with AI. Things have stalled here and in the European Union. The EU though has plans to roll out its rules on content labelling next year. There’s no date on the UK side. Platforms like Midjourney though do include labelling technology. You can embed invisible metadata and watermarks into generated images to signal they’re AI-generated. There are also initiatives to use nutrition-like labels for digital content. Users can embed the origins and editing history of the AI creative. Brands like Adobe are out in front on this. Learn more here.

We’re open with clients about when and where AI is used in their work. It’s a conversation worth having at the start of every project – to understand your appetite for AI, set expectations, and make sure everyone’s comfortable with how it’s being used.”

How do we protect data?

Use paid, enterprise-grade tools – tools that guarantee zero data retention and no model training. Uploading brand guidelines, for example, to a public tool is fraught with risks, from generating assets that infringe another brand’s IP to data leaks. If you’re unsure defer to an expert.

Customer data – comply with GDPR – not much else we can say here.

Anonymise sensitive information – this is pretty straightforward.

Platform Ts&Cs – as mentioned earlier check policies on how third-party trademarks and data are handled and ownership rights are allocated.

Protect your website content: Update your robots.txt file / block AI scrapers from your website – it’s not foolproof, but it sets a clear boundary and helps establish intent if you ever need to enforce your rights. There are services that can do the heavy lifting for you. Cloudflare is the most accessible – it now blocks AI scrapers by default and has a one-click toggle to block all known AI crawlers, including GPTBot, ClaudeBot, Meta and ByteDance. It’s available on their free plan, so there’s no reason not to. They’re also rolling out a pay-per-crawl model, meaning AI companies would need to pay to access your content rather than taking it for free. Worth knowing: blocking scrapers protects your content but may reduce how often your brand appears in AI-generated answers and recommendations – it’s a trade-off worth discussing with your marketing team.  WARNING! The big trade off here is that blocking AI scrapers protects your content but may reduce how often your brand appears in AI-generated answers and recommendations. It’s a trade-off worth discussing with your marketing team.

Opt out of AI training on platforms: If you have an in-house creative team, audit the tools they’re using – Adobe, Canva, Microsoft Copilot, Google Workspace – review the data and training settings, opt out wherever you can, and remove any content you don’t want scraped.”

[Note to Angela: could you add a couple of lines on what you do to opt out of scraping and how you’ve updated your website Ts&Cs?]

A last word

The law is still catching up with AI, but that doesn’t mean you have to wait. The agencies and brands getting ahead are the ones having these conversations now.

Want to start that conversation – let’s schedule some time together. Connect with us here.