MLex: Large tech companies are providing EU enforcers with details on how many people use their AI assistants in a move that could lead to tighter regulation under digital gatekeeper rules, MLex has learned. Assistants run by the likes of Alphabet, Anthropic, Meta Platforms, Microsoft and OpenAI are increasingly common, and regulators have faced pressure over whether they should be regulated under the EU Digital Markets Act (EU DMA) as key gateways for reaching users.

Artificial Intelligence (AI) companies such as OpenAI and Google DeepMind may soon need to secure licensing agreements with European rights holders to use content protected by copyright for training generative AI applications like ChatGPT and Gemini. The European Commission aims to develop a licensing market for training data by introducing a robust reservation rights mechanism, enabling rights holders to protect their content from data scraping.

MLex: The European Commission’s AI Office is analysing whether there are any signs that AI companions allow illegal child sexual abuse material or non-intentional intimate images, a Commission Official has said. Under the EU AI Act, AI companions are treated as high concern use cases where the Commission will test and enforce the new prohibitions and general safety obligations, especially when those companions are built on general-purpose AI models with systemic risk.

This resource kit contains a list of the key practical guidance available across Lexis+® UK that deals with issues that arise when procuring artificial intelligence (AI) technologies and negotiating AI contracts. It is updated as new content is added. For broader content relating to AI technologies and the legal issues arising from its use in different contexts and sectors, see Practice Note: Artificial intelligence (AI) resource kit. AI procurement and contract resources Guides, checklists and overviews Practical guidance document More information Practical guidance content type Negotiation guide—AI contracts This negotiation guide sets out practical guidance on some of the key issues that may arise when negotiating contracts to procure AI technologies. It is primarily designed to support customers procuring AI solutions but it considers some of the positions that suppliers may take. It summarises the aims of the parties, considers possible negotiating positions and highlights common pitfalls that can be encountered when negotiating AI contracts. Practice Note The AI project lifecycle—a quick guide Realising value from AI while managing the risks involves

MLex: AI developers and copyright holders will see the government’s initiative to resolve a stand-off over the use of copyrighted materials in training artificial intelligence ‘by the end of the year’, the UK’s Parliamentary Under-Secretary of State for Science, Innovation & Technology, Feryal Clark said. She stressed that ‘whether that is [through] legislation or amendment to a policy, that is yet to be decided’, but she still expected a resolution ‘in the very near future, by the end of this year’.

MLex: AI model providers and right holders are seeing the AI-related commitments in the EU Code of Practice for general-purpose AI (GPAI) further watered down. A new iteration of the text, released on 11 March 2025, makes the commitments proportionate to the size of the provider, removes the requirements to verify the copyright compliance of third-party datasets and limits the possibility to lodge complaints, among other things. The latest draft also introduces significant changes to the transparency requirements.

General-purpose AI model providers like Microsoft, OpenAI and Google will now see softened copyright-related measures in the second draft of the EU’s AI Code of Practice.

Law360, London: Artificial intelligence (AI) will be a 'game changer' for the reinsurance industry as it navigates an increasingly turbulent geopolitical landscape, Swiss Re said in a report 8 September 2025.

MLex: Developers and users of artificial intelligence (AI) systems will be able to use voluntary guidance on how to apply an upcoming EU obligation on AI literacy by the end of 2025, European Commission officials told national governments at a closed-door meeting on 10 December 2024.

MLex: OpenAI, Microsoft, Google and other big AI companies saw the final version of the EU AI Act’s code of practice for general-purpose AI models published today, although guidelines and a data disclosure template are still pending, and it remains unclear whether AI companies will commit without the full picture. The European Commission is informally discussing a potential grace period to ease initial compliance, but its length is still to be confirmed. Some final changes made it into the last version of the text, with mixed results on a crucial part on copyright.

MLex: Draft procedural rules for enforcing the EU AI Act would give the European Commission the power to access and inspect general-purpose AI models, including source code, model weights and hosting infrastructure. The plans also set out independence criteria for experts, introduces five-year limitation periods for fines and enforcement, and would guarantee providers at least 14 days to respond to preliminary findings.

MLex: The European Commission published on 18 July 2025 its guidelines on the EU AI Act’s rules for general-purpose AI (GPAI) models, setting out its interpretation of the provisions’ scope, key concepts and application. The document outlines compute thresholds for defining a model as GPAI, and for when a modification constitutes the creation of a new model. It also explains when a model is considered as placed in the EU market, which models will fall under the open-source exception and how the requirements will be applied to models already on the market.