Share this article:

AI Observer | Series 2, Issue 1: Agreements for AI services, Google’s €250 million copyright fine and what’s happening in the UK AI legal landscape

By Louise Perkin, Partner and Caroline Day, Partner and Gemma Robin, Partner and Stephan Schultes, Senior Associate

Welcome back to the AI Observer, HLK’s newsletter tracking the AI legal landscape.

Our team of AI experts kick off Series 2 by delving into the differences between traditional IT agreements and agreements relating to the use of AI and update you on some key legal developments in the UK and Europe.

Did you miss Series 1 of the AI Observer? Check out all our previous issues and other AI updates here.

Navigating the differences between AI agreements and more traditional IT agreements

IT agreements are used to set out certain expectations in relation to delivery of IT services and products. The purpose of such an agreement is to provide clarity and certainty, thereby fostering an effective and positive business relationship.

So, what are the differences between the terms of AI agreements and those of traditional IT agreements?

Louise Perkin and Sophie Harrison delve into the answer, which is not straightforward because much depends on the specific AI and IT tools and the point in time at which the agreement is being negotiated (legislative changes, industry preferences, etc.).

Read the full article here.

Legislation watch

In big news for those following the extended wranglings over use of copyright material in training data for gen AI, March saw Google become the first AI company to be fined at least in part over the use of copyright material in training data. This is part of a long running dispute between the tech giant and the French Competition Authority.

Back in 2022, Google made commitments about entering good faith negotiations with French news outlets about compensation for including article extracts in search results. Google was found to have failed to respect those commitments, and, most interestingly for our purposes, the fine print is stated to take account of the fact that Google “used content from press agencies and publishers to train its foundation model, without notifying either them or [the French Competition Authority]”.

The foundation model in question is Google’s Bard AI, now known as Gemini. This appears to be the first time that a company has been fined for the unauthorised use of copyright material in training data. In addition, with so much focus on US lawsuits for copyright infringement, it is interesting to note that the French Competition Authority indicated that the question of whether or not article-scraping for AI training data qualifies for the text and data mining exception of the EU Copyright Directive “has not yet been settled”. Perhaps not surprising then that OpenAI has started making licensing agreements with European press publishers including Axel Springer and Le Monde. 

On the regulatory front, it appears the UK may be rethinking its position on AI regulation in light of concern over risks, and, as time ticks down to the first deadline for applicability of the EU AI Act, we may be seeing the first moves by large players to ensure compliance. Meta recently updated its policy for content labelling, with much more content being labelled as generated or manipulated (“imagined”) using AI. This move follows a lot of consultation, but also helpfully seems to be broadly in line with requirements of the AI act for labelling of material generated with AI.   

Are neural networks patentable in the UK? An appeal hearing in the hotly watched Emotional Perception AI vs UK Patent office last week has taken us one step closer to finding out. The UKIPO appealed a previous Decision of the High Court which conclude that a trained ANN is not excluded from patentability as a computer program. The decision should issue within a few weeks.

AI application of the month

If there’s one thing worse than stepping barefoot on a pile of Lego® bricks, it’s having to search through them to find that one brick you need to complete your brick-built creation!

But help has arrived in the form of AI-powered apps, such as Brickit, that can scan a pile of bricks using the mobile phone camera, recognise individual bricks, and even pinpoint their location to the user. So get building!




This is for general information only and does not constitute legal advice. Should you require advice on this or any other topic then please contact or your usual HLK advisor.

HLK bubble graphic HLK bubble graphic

Join the #ImagineBetterAI conversation on our social media channels. You can also sign up to receive our monthly AI Observer newsletter here.