Featured documents
R (Greyhound Board of Great Britain Ltd) v Welsh Ministers [2026] EWHC 670 (Admin) What are the practical implications of this case? The ruling reinforces the constitutional divide between the courts and the legislature. It explains that the scheme and framework of the Government of Wales Act 2006 (GWA 2006) embody that separation of powers, and that any judicial attempt to recognise and enforce a common law obligation on Welsh Ministers to consult prior to introducing legislation in the Senedd would trespass upon that boundary. This is not a departure from established principle; case law has already upheld comparable rules for lawmakers in Scotland and at Westminster. However, this is the first express confirmation of the position for Welsh lawmakers, and the first time this dimension of the GWA 2006 has been analysed in such depth. The court examined earlier
The solution arrived through the United Nations Compensation Commission (UNCC), a quasi‑judicial body handling mass claims, created under UN Security Council Resolution 687. By addressing environmental harm—most notably via its ‘F4’ claim class—the UNCC set a seminal benchmark shaping how international law and contemporary arbitral panels allocate financial responsibility for wartime ecological devastation. With present-day wars in areas such as Eastern Europe and the Middle East bringing dam breaches, strikes on chemical facilities, and the burning of farmland, the UNCC’s legacy endures as an essential reference point for states, global investors, and companies engaged in post‑conflict arbitration. The F4 claims: Quantifying the unquantifiable Prior to the 1990s, mechanisms in international law for war reparations overwhelmingly favoured property loss, foregone earnings, and bodily injury. The natural world was commonly treated as a mute, non-compensable victim of armed hostilities...
Understanding the farming business as a business Many farms still use long-standing structures that arose by habit, not strategy. Sole traders, informal partnerships and outdated partnership deeds are common. While once effective, such setups can cause major issues around succession, tax planning and involving the next generation. A corporate team can take a fresh, business-led view of the farm, asking: Who owns the land and other critical assets? Who manages daily operations? Who carries the risk and who enjoys the return? What is the enduring plan for succession? From this review, the team can confirm whether the current setup is fit for purpose or if an alternative — for example an updated partnership agreement, a company, a limited liability partnership, or a blended model — would better meet the family’s aims. Tax efficiency through joined-up advice Tax sits at the centre of most
Most recent News
The window for submissions to the Labour government’s consultation closed on 25 February 2025, and practitioners warn the reforms face an all-or-nothing outcome. Paul Joseph, a partner at Linklaters LLP, said he suspects the choice is either to retain the current position or to adopt the EU model; his take on the copyright consultation is that there appears to be sufficient momentum to follow the EU approach, though only time will reveal the result. Opening in December 2024, the consultation proposed an exception to data-mining rules for businesses training AI models for commercial use (see: LNB News 19/12/2024 6). It also moots giving creatives an opt-out from such training in an effort to balance their rights. In essence, the plan would broaden text and data-mining exceptions to align with the EU, which is working to implement a similar opt-out for copyright works after...
In this issue: New technologies Information technology Internet Advertising, marketing and sponsorship Lex Talk®TMT: a Lexis®Nexis community Daily and weekly news alerts New and updated content Dates for your diary Trackers Useful information New technologies AI in dispute resolution—balancing the risks and opportunities Dispute Resolution analysis: AI stands to reshape dispute resolution, boosting efficiency and cutting costs. As AI’s role within the legal sphere accelerates, there is a growing obligation to balance its promise against likely risks to reliability, accountability, privacy and ethics. This article examines the changing place of AI in dispute resolution, the advantages it can bring, the drawbacks it presents, and the importance of planning to reduce the potential hazards it may create. Authored by Rebecca Warder, Head of Knowledge Management, Hausfeld & Co, Lucy Pert, Co- Head of Commercial Disputes and...
What is the Deepseek AI model? Deep Seek AI is a cutting-edge artificial intelligence system built by Deep Seek, a company based in China. The organisation’s mission centres on stretching the limits of AI to deliver tools that empower enterprises and elevate human capability, while keeping advanced AI accessible, ethical and meaningful. Established in 2023, Deep Seek rose swiftly through the AI ranks by releasing free, open-source language models. Most recently it unveiled two high-end models: V3, aimed at broad use cases such as chat-style applications, and R1, tailored for reasoning-heavy work, including programming and maths challenges. R1 has drawn notable media interest by offering cost-conscious AI performance when set against leading US counterparts. Why is it having a big impact on technology markets around the world? Deep Seek AI is reshaping global tech markets largely through its cost efficiency. Reports suggest the company trained its system for...
The EU AI Act adopts a layered regime for AI systems, imposing transparency duties on every model and reserving stricter risk‑management obligations for the most potent ‘systemic’ ones. These obligations are being hammered out through a code of practice intended as a key compliance tool for model providers. Yet that code could prove marginal if models are not correctly classified for systemic risk Under the Act, a model is chiefly deemed systemic when its training consumed more than 10^25 floating‑point operations. According to research institute Epoch AI, several flagship systems have already crossed this line, including: Open AI’s o1 Anthropic’s Claude 3.5 Sonnet Mistral AI’s Large 2 These are commonly viewed as ‘frontier’ models, pushing the limits of this disruptive technology. Nonetheless, recent developments increasingly call into question whether the volume of compute used in the pre‑training stage is the most...
Article 45 of the DSA Article 45 of the DSA provides that, when significant systemic risks arise under Article 34(1) (which sets out the duty on very large online platforms ( VLOPs) and very large online search engines ( VLOSEs) to identify, analyse, and assess systemic risks), and affect multiple VLOPs or VLOSEs, the Commission may ask those VLOPs and VLOSEs to help draft codes of conduct, with commitments to adopt risk mitigation measures and to report on those measures and their results. The Code of Conduct+ was adopted in this context. VLOPs’ and VLOSEs’ adherence to the Code of Conduct+ can count as a mitigation measure under Article 35 of the DSA; however, participation in and implementation of the Code of Conduct+ does not, by itself, amount to compliance with the DSA ( Recital 104)......
In this issue: New technologies Data protection Media Advertising, marketing and sponsorship Telecommunications Lex Talk®TMT: a Lexis®Nexis community Daily and weekly news alerts New and updated content Dates for your diary Trackers Useful information New technologies European Commission issues guidelines on the definition of AI systems The European Commission has released long‑anticipated, detailed guidance on the definition of artificial intelligence ( AI) systems set by Regulation ( EU) 2024/1689 ( EU AI Act). This practical guidance is designed to help providers, deployers, importers and distributors judge whether a system qualifies as an AI system under the EU AI Act, supporting consistent application and enforcement. The definition of AI systems has applied since 2 February 2025. Dóra Petrányi, Central Eastern Europe managing director and co‑head of the Technology, Media and...
Key elements of the AI system definition Given the wide variety of AI systems, the guidance cannot offer a complete list. Each system must be evaluated on its own particular characteristics. The EU AI Act adopts a lifecycle-oriented approach and characterises an AI system as: a machine-based system designed to operate with differing degrees of autonomy that may show adaptability after deployment and, for explicit or implicit aims infers from the inputs it receives how to generate outputs such as predictions, content, recommendations, or decisions that can affect physical or virtual environments This seven-part definition spans two main stages: pre-deployment (building phase) and post-deployment (use phase). The seven elements need not be present throughout both stages; some may appear only at one point. The following offers a brief summary of each of the seven points listed...
The publishers, including Guardian News & Media Ltd, the Atlantic Monthly Group LLC and the Mc Clatchy Co LLC, have filed a complaint accusing the Canadian AI firm of using copyright-protected works to train its large language models ( LLMs). It is the latest action by media companies and publishers against LLM developers that typically rely on vast bodies of data. The claimants contend Cohere went beyond training on restricted material: its systems sometimes reproduce protected content word for word, or generate incorrect answers that are wrongly attributed to the publishers. The complaint asserts the company has not only taken their works but also blatantly invents fake pieces and links them to the publishers, misleading the public and damaging their brands... A Cohere spokesperson said the company strongly stands by its practices for responsibly training its enterprise AI, adding that it has long...
Macdonald Hotels Ltd v Bank of Scotland Plc [2025] EWHC 32 ( Comm) Relevant facts The High Court examined allegations brought against Bank of Scotland Plc ( BOS) by Macdonald Hotels Limited ( MHL) arising from the compelled sale of three properties: the Randolph Hotel, the Old England Hotel and the Marine Hotel (the Hotels). MHL contended that disposals occurred when valuations were unusually depressed, contravening implied terms of a facility agreement and, in relation to the Randolph, express provisions in a shareholders agreement. MHL further maintained that BOS ought to have allowed repayment by alternative means (including third‑party refinancing) and/or over an extended timescale. Resolving the dispute required review of years of dialogue, intricate documentation, and numerous detailed amendments. Careful consideration was required to identify and fully determine the issues arising in the case. This note concentrates on the funding...
ICO response to Google’s device fingerprinting decision In December 2024, the ICO responded to Google’s move to lift its ban on device fingerprinting (gathering and merging details about a device’s software and hardware to identify it) for organisations using its advertising products, with the change effective from 16 February 2025. See: LNB News 19/12/2024 73. This followed Google’s July 2024 decision to keep third party cookies. In its response, the ICO condemned Google’s choice to allow device fingerprinting for advertising as ‘irresponsible’ and underlined that device fingerprinting: requires consent — it can identify devices even when cookies are blocked or location is hidden; although commonly deployed for fraud prevention, the ICO reiterated it remains subject to the usual consent rules reduces user control — despite browsers now offering ‘enhanced’ tracking protection, the ICO said fingerprinting is not a fair way to track people online as it...
Getty Images ( Us), Inc and other companies v Stability Al Ltd [2025] EWHC 38 ( Ch) What are the practical implications of this case? The decision highlights three practical points when pursuing claims in a representative capacity. Parties must understand the criteria for representative actions (succinctly set out at paragraph [57] of the judgment). In particular: define the class under CPR 19.8 precisely so every member can be identified ensure the class definition is not contingent on the litigation outcome; avoid definitions tied to disputed issues all class members must share the same interest; beware framing issues at too general a level, which unravels on closer analysis Sensible case management is equally vital, especially in complex disputes. The court repeatedly indicated—adversely to the representative route—that permitting such a claim without clear, worked‑through proposals for how the claim(s) would be dealt with would not sit...
What are the practical implications of this case? EE Ltd versus Virgin Mobile Telecoms Ltd [2025] EWCA Civ 70...
The EU AI Liability Directive The EU AI Liability Directive surfaced on the Commission executive’s list of files to be scrapped at the eleventh hour, following pressure from EU digital chief, Henna Virkkunen, and as part of a wider push to reduce and simplify the regulatory framework. At a press briefing on the Commission’s 2025 work programme (which outlines upcoming initiatives and those to be abandoned), Šefčovič, responsible for deepening the Commission’s relations with lawmakers and Member States, said Commissioners had been examining how files marked for withdrawal were advancing through the co‑legislative process. ‘ When we observe that these particular proposals are stuck, at times for many years, we harbour serious doubts that they will progress this year,’ he said. The EU AI Liability Directive, a plan to harmonise certain administrative aspects of AI‑related damage claim proceedings, has indeed advanced little since it was...
In this issue: New technologies Internet Data protection Media Advertising, marketing and sponsorship Reputation management Telecommunications Lex Talk®TMT: a Lexis®Nexis community Daily and weekly news alerts New and updated content Dates for your diary Trackers Useful information New technologies European Commission issues guidance on AI system definition The Commission has issued guidance to help providers judge whether a software system qualifies as artificial intelligence under Regulation ( EU) 2024/1689 (the EU AI Act). Though not binding, the material is intended to support application of the Act’s provisions, which started on 2 February 2025. The Commission notes the guidance will be refined over time in light of real-world experience and emerging scenarios. Under the Act, AI systems are grouped by risk—prohibited, high risk, and those carrying transparency duties—and this guidance sits alongside existing advice on...
The Code joins an expanding roster of UK cybersecurity measures, including: the Code of Practice for Software Vendors the Cyber Governance Code of Practice the Consumer Io T Code of Practice the Product Security and Telecommunications Infrastructure Act 2022 the Code of Practice for app store operators and app developers It also sits within a broader UK cybersecurity strategy that features a new UK Cyber Security and Resilience Bill and likely updates to the Network and Information Systems Regulations 2018, with the prospect of further legislation—for example, the recent consultation on ransomware payments. Background to the Code The UK Government acknowledged that software must be secure by design (a term familiar to those versed in data protection law) and that stakeholders across the AI supply chain need certainty on baseline security expectations. Developed by the Department for Science,...
The ICO stresses it has acted swiftly in step with rapid advances in generative AI. Demonstrating this agility, it opened a consultation series in January 2024 focused on generative AI and data protection. Its aim was to set out how organisations might build and implement generative AI while meeting UK data protection duties, especially those in the United Kingdom General Data Protection Regulation, Assimilated Regulation ( EU) 2016/679 ( UK GDPR). The ICO highlighted that adherence to the data protection framework is paramount when using generative AI, as such models are commonly trained on vast volumes of personal data. The consultation raised worries about insufficient transparency over how personal information is used within generative AI, which in turn creates the risk that data protection rights could be undermined......
In this issue: New technologies Internet Data protection Media Advertising, marketing and sponsorship Reputation management Telecommunications Lex Talk®TMT: a Lexis®Nexis community Daily and weekly news alerts New and updated content Dates for your diary Trackers Useful information New technologies An in-depth look at the EDPB Opinion on personal data processing within AI models On 17 December 2024, the European Data Protection Board ( EDPB) issued Opinion 28/2024, addressing specific data protection questions arising from the processing of personal data in the sphere of artificial intelligence ( AI) models. The Opinion emphasises a firm expectation that controllers rigorously evaluate and document their determination of an AI model’s anonymity, including an appraisal of the likelihood of identification, measured against the criteria set out in the Opinion. Alex Jameson of Bird & Bird examines the...
The Opinion arose from a referral by the Irish Data Protection Commission ( DPC) under Article 64(2) of the EU General Data Protection Regulation, Regulation ( EU) 2016/679 (the EU GDPR). Article 62(4) allows any EU data protection authority ( DPA) to put questions to the EDPB that are of general relevance or have implications across more than one EU Member State, with the aim of enhancing harmonisation and clarity on such matters. Over the past twelve months in particular, DPAs have been wrestling with AI, and the DPC—acting as lead DPA for many of the world’s largest technology companies—has frequently led on the most urgent issues. While EDPB Opinions, akin to guidelines, do not have binding legal force, they will strongly shape how DPAs proceed; accordingly, we should expect the Opinion’s principles to appear in national guidance and...
On 23 January 2025, the CMA unveiled two further investigations into whether any company possesses SMS in relation to certain digital activities, this time concentrated on Apple and Google’s mobile ecosystems (see LNB News 23/01/2025 15). The CMA will evaluate whether each of Apple and Google has SMS for the provision of specified services within their mobile ecosystems and, if so, whether to impose measures (conduct requirements) to protect competition, consumers and/or businesses. The announcement lands just a week after the CMA disclosed its first SMS investigation, concerning Google’s general search services (see LNB News 14/01/2025 26). The emergence of three investigations in quick succession shows the CMA moving promptly to deploy the strengthened powers to review digital markets that it obtained at the start of 2025. What is a ‘mobile ecosystem? ......
What is the CNI regime? Critical National Infrastructure ( CNI) encompasses the physical and digital assets, sites, and networks that keep a nation running, underpinning the health, safety, security, and economic prosperity of its people. In the UK, the basis for CNI rests on a mix of long-standing government programmes, statutory duties, and policy directives intended to protect pivotal industries. Together, they support the functioning of the country and the well-being of its population. Collectively, these elements are indispensable to national operations and the welfare of citizens. The regime arose from the imperative to shield vital services and infrastructure central to national security, economic resilience, and public protection—a subject under review by many states from the mid-twentieth century. That imperative sharpened towards the close of the twentieth century as threats from terrorism, espionage, and cyberattacks...
Popular documents
When evaluating a general damages claim, the practitioner ought initially to refer to the Judicial College Guidelines (JCG)...
This Practice Note This Practice Note reviews mechanisms used in settling litigation. A Tomlin order consists of a consent order paired with a schedule. It operates to stay proceedings on terms that have been agreed. The provisions contained in the schedule may remain confidential. This Practice Note describes the scope of confidentiality attaching to the schedule and sets out how it differs from a standard consent order. Sample wording for a Tomlin order is included, alongside links to precedents, as well as guidance on court approval. It also addresses varying, setting aside and enforcing a Tomlin order, including the considerations the court will take into account when handling applications for each. Further guidance is provided on interpreting and applying the relevant provisions of the CPR; however, some courts and divisions impose very specific requirements for both drafting and approval, and for approaching the schedule and confidentiality issues. Accordingly, you must consider the particular rules and court guide provisions in the forum where your claim is proceeding when drawing up the Tomlin order...
Date [ date ] Parties [ name of Landlord ] [ of OR incorporated in England and Wales (company registration number [ number ]) with its registered office at ] [ address ] (Landlord) [ name of Tenant ] [ of OR incorporated in England and Wales (company registration number [ number ]) with its registered office at ] [ address ] (Tenant) [ [ name of Guarantor ] [ of OR incorporated in England and Wales (company registration number [ number ]) with its registered office at ] [ address ] (Guarantor) ] [ [ name of Mortgagee ] [ of OR incorporated in England and Wales (company registration number [ number ]) with its registered office at ] [ address ] (Mortgagee) ] Definitions Within this Deed, the terms below shall be interpreted as follows: [ Annual Rent • the annual sum reserved under the Lease; ] [ Insurance Rent • the Tenant’s share of the Landlord’s costs of insuring the Property (as set out in the Lease); ] Lease • the lease of the Property dated [ date ], entered into between (1) [ the Landlord OR [ name ...
I, [ name ], of [ address ], solemnly and sincerely state that: [ Matters to be verified, set out in numbered paragraphs ] I make this solemn statement in good conscience, believing it to be true, and pursuant to the provisions of the Statutory Declarations Act 1835. DECLARED at [ details ] this [ day ] day of [ month and year ] Before me ................................................................................ [ signature of the person before whom the declaration is made ] A [ commissioner for oaths OR [ solicitor OR [ insert other qualification ] ] authorised to administer oaths ]...
Discover more from LexisNexis
