As we wrap up another eventful month in the world of AI, we're excited to share some of the key developments that have caught our eye. On the new model front, we saw a number of new releases, including DeepSeek V3-0324, Google DeepMind’s Gemini 2.5, Alibaba’s Qwen2.5-Omni-7B and Tencent's T1 Reasoning Model, to name a few, as well as the launch of the Manus AI agent from Chinese start-up Monica. Developed with backing from Tencent Holdings, Manus deploys a hierarchical system of specialised AI agents co-ordinated by a central "task master" which claims to independently handle multi-step workflows, such as analysing CVs, creating websites or conducting research studies. Many expect agentic AI to transform industries, with NVIDIA's CEO Jensen Huang, fresh from this year’s GTC, envisaging a future where employees oversee millions of AI agents to optimise workflows. March also saw a number of significant policy initiatives and regulatory events.
UK initiatives
UK copyright and AI: attempting to balance innovation and creative rights
The UK government is actively exploring reforms to copyright law to address the growing influence of AI technologies, particularly in the realm of AI-generated art and training models. Key proposals include introducing a copyright exception for AI training, allowing developers to use copyrighted works without explicit permission, provided rights holders can opt out. This has sparked significant debate, with creators expressing concerns over diminished control and compensation for their works, while AI developers emphasise the necessity of access to high-quality materials for innovation. The consultation process, which recently concluded, aims to strike a balance between fostering AI development and safeguarding creative industries, but uncertainty remains over ownership and protection of AI-generated content. Whilst the government has pledged to address creatives’ concerns through transparency measures and collaboration between AI developers and content creators, the debate highlights tensions between fostering innovation and safeguarding intellectual property rights.
UK CMA roadmap for big tech regulation
The Competition and Markets Authority (CMA) has outlined its strategic plan to regulate major tech firms, focusing on interventions that directly impact UK businesses and consumers. Using new powers under the Digital Markets, Competition and Consumers Act, the CMA will investigate companies with "Strategic Market Status" (SMS), such as Google and Apple, to ensure fair competition and prevent anti-competitive practices. The roadmap, set for release in mid-2025, will prioritise transparency, proportionality and fostering innovation while addressing concerns about digital monopolies. The CMA’s CEO, Sarah Cardell, emphasised the importance of balancing regulation with economic growth to maintain a competitive and fair digital marketplace.
UK government adopts ‘test and learn’ approach for AI innovation
The UK government has unveiled a new "test and learn" strategy to fund AI and digital projects, aiming to reduce waste and accelerate innovation in public services. This approach involves allocating small budgets to experimental initiatives, with additional funding granted to successful prototypes. Inspired by start-up practices, the reforms seek to streamline bureaucratic processes and modernise public sector operations. From April, the government will trial four funding models, including staged funding, to ensure promising technologies are scaled effectively while minimising risks.
UK watchdog ends probe into Microsoft’s OpenAI investment
The CMA has closed its investigation into Microsoft’s $13 billion investment in OpenAI, ruling that the partnership does not qualify as a merger under UK law. While acknowledging Microsoft’s “material influence” over OpenAI, the CMA found no evidence of de facto control that would warrant further scrutiny. Critics, however, have raised concerns about regulatory leniency amid government pressure to prioritise economic growth. Microsoft welcomed the decision, emphasising its commitment to innovation and competition in AI development.
Oracle to invest $5 billion to boost UK AI innovation
Oracle announced a $5 billion investment over the next five years to expand its cloud infrastructure in the UK, supporting the government’s ambition to become a global AI leader. The funding will enhance Oracle Cloud Infrastructure’s footprint, enabling businesses and public sector organisations across industries like healthcare, defence and manufacturing to leverage cutting-edge AI technologies. This move aligns with the UK’s AI strategy and aims to drive productivity, innovation and economic growth while cementing the nation’s position at the forefront of the AI revolution.
European initiatives
EU AI Code of Practice: third draft dilutes protections for creators
The third draft of the EU General-Purpose AI Code of Practice appears to have taken a lighter touch than the prior drafts in the field of copyright compliance. While the earlier drafts emphasised stringent requirements for copyright due diligence and anti-piracy measures, the latest version introduces more flexible language and eases some obligations, including reduced transparency requirements and flexibility in copyright due diligence, acknowledging the difficulty of verifying the copyright status of large datasets, permitting developers to rely on "credible sources" without exhaustive checks. Instead of imposing strict penalties for non-compliance, the draft’s focus has shifted towards collaborative measures, encouraging AI providers to work with rights holders to resolve disputes.
Potential reopening of EU AI Act for SME compliance simplification
Lucilla Sioli, Director of the European Commission’s AI Office, has confirmed that the AI Act will be included in the Commission's upcoming digital package, with an impact assessment expected by late 2025. Speaking at a privacy panel, Sioli clarified that the focus will be on simplifying compliance for SMEs rather than conducting a "big review" of the Act. The assessment will likely examine intersections with other legislation to streamline obligations for smaller businesses.
Tensions emerge between EU AI Act and GDPR over algorithmic discrimination
Several experts, including Stefano De Luca from the European Parliamentary Research Service and Michael Will, President of the Bavarian State Office for Data Protection Supervision, have suggested that the EU's AI Act and the General Data Protection Regulation (GDPR) are creating legal uncertainty due to their differing approaches to algorithmic discrimination. The AI Act allows processing of sensitive personal data to mitigate bias in high-risk AI systems, but the GDPR imposes stricter limits on such data processing. This tension raises questions about how to reconcile these regulations, potentially requiring legislative reform or additional guidelines to clarify their interplay and ensure compliance with both AI innovation and data protection principles.
EU establishes scientific panel to support AI Act enforcement
The European Commission has adopted an implementing act to establish a scientific advisory panel of independent AI experts, as mandated by the AI Act. This panel will assist the AI Office and national authorities by providing technical advice on enforcement, developing assessment tools and issuing alerts about risks from general-purpose AI models. The Act outlines the panel's operational rules, with a call for expressions of interest to select experts expected soon. This initiative aims to strengthen governance and ensure effective implementation of the AI Act across the EU.
Elsewhere…
Chegg sues Google over AI summaries, citing antitrust violations
Educational technology company Chegg has filed a landmark antitrust lawsuit against Google, alleging that its AI-powered search summaries, known as AI Overviews, have severely impacted traffic to its platform and undermined the broader digital publishing ecosystem. Chegg claims Google uses its dominant position in search to repurpose proprietary content without proper attribution or compensation, keeping users within Google's ecosystem and reducing incentives for publishers to create original content. The lawsuit highlights concerns about monopolistic practices in AI-driven search and raises broader questions about the future of online information access. Chegg filed its lawsuit against Google in the District Court for the District of Columbia, where it has been assigned to federal judge Amit Mehta.
Palantir delivers first AI-enabled TITAN systems to US army
Palantir Technologies has delivered its first two AI-powered Tactical Intelligence Targeting Access Node (TITAN) systems to the US army. These mobile ground stations use AI to process satellite and sensor data, enabling autonomous decision-making without cloud connectivity, and follow the delivery of the initial prototype in August 2024, as part of a larger contract aimed at equipping the army with next-generation intelligence systems. Designed for real-time intelligence, TITAN systems consist of advanced and basic variants mounted on tactical vehicles. The army is evaluating the initial systems delivered with the potential to procure up to 150 units for broader deployment by 2026.
UiPath acquires Peak AI to boost vertical AI solutions
UiPath, a leading enterprise automation and AI software company, has acquired Peak AI, a Manchester-based AI specialist known for its decision-making AI solutions in inventory and pricing management. This strategic move aims to enhance UiPath’s vertical AI solutions strategy by integrating Peak’s purpose-built AI applications, which optimise business processes across industries like retail and manufacturing. The acquisition enables Peak to scale its solutions globally, leveraging UiPath’s extensive reach and expertise to deliver next-generation AI-driven applications powered by LLMs and agentic automation.
CoreWeave IPO: a mixed debut for an AI sector bellwether
CoreWeave, an AI-focused cloud computing provider, launched its IPO on Nasdaq on 28 March, raising $1.5 billion - below its initial $4 billion target. Shares closed flat at $40, reflecting mixed investor sentiment amid broader tech market struggles. The company aims to capitalise on the growing demand for AI infrastructure, with major clients like Microsoft and NVIDIA driving 77% of its revenue. While its entry marks the largest US tech IPO since 2021 and a potential bellwether for the AI sector, concerns over customer concentration, high debt levels and the strength of its moat, cast uncertainty on its long-term outlook. Some industry analysts view this debut as a critical test for generative AI investment trends.
We are seeing a growing focus on enterprise applications of AI across the industry, with companies developing AI platforms that prioritise optimised performance, profitability and security to meet the needs of their enterprise customers. Some key trends include AI reasoning and custom silicon fuelling demand for chips, cloud migrations and AI workloads as revenue opportunities for hyperscalers, data companies focusing on evaluating AI efficacy, and software companies setting sights on agentic AI.
As the AI landscape continues to evolve at a rapid pace, we aim to stay abreast of the latest developments and their potential implications for businesses and legal practices. Wherever you are on your AI journey, we would love to hear from you.
Get in touch with Tim Wright or Nathan Evans if you would like to discuss any of the contents of this article in more detail.
