AI action plan

In January the government published the independently written AI Opportunities Action Plan and pledged to take forward all 50 of its recommendations. Key points include the following sections:

1.2 Unlocking data assets in the public and private sector

Arguably building on the UK Data (Use and Access) Bill covered in December 2024’s What’s New, this part of the plan seeks to “unlock both public and private data sets to enable innovation by UK startups and researchers and to attract international talent and capital”. It will include creating a National Data Library (NDL) to establish a “copyright-cleared British media asset training data set”.

1.4 Enabling safe and trusted AI development and adoption through regulation, safety and assurance

This seeks to balance the regulation of AI and supporting the AI Safety Institute (AISI) alongside other regulators, whilst at the same time ensuring that the regulatory environment is attractive for entrepreneurs and investors. It includes a rather controversial point (24) which is to: “Reform the UK text and data mining regime so that it is at least as competitive as the EU.” This should be read in conjunction with the Copyright and Artificial Intelligence consultation (see below).

According to Prime Minister Keir Starmer, the AI Action Plan sets out a roadmap to the UK becoming “an AI superpower”, and newly appointed AI government adviser Sir Demis Hassabis (co-founder of Google Deepmind) said it will “help the UK unleash AI’s potential to drive growth, accelerate scientific discovery and tackle important, real-world problems.”

It’s also worth noting the recent publication of a “blueprint for modern digital government“ which sets out a six-point plan for government digital reform, the second of which is to “Harness the power of AI for the public good”.

Copyright and artificial intelligence consultation

Currently there is a lot of uncertainty around the legality of training LLMs on copyright data. A consultation ran from December 2024 to February 2025, seeking views on “how the government can ensure the UK’s legal framework for AI and copyright supports the UK creative industries and AI sector together”. One of the core – and highly divisive – questions in the consultation, is the proposed introduction of an exception to copyright law for text and data mining (TDM), which would essentially allow AI providers to train their LLMs without fear of being sued for copyright infringement. This would bring the UK in line with the EU’s TDM exception contained in its Digital Single Market (DSM) Directive 2019/790. 

More than 1,000 artists, including Kate Bush and Damon Albarn, have released a silent album in protest at the TDM exception plans, and the Daily Mail is running a campaign urging the government to change course.

Also relevant to this debate are the proposed amendments to the Data (Use and Access) Bill, which will require AI companies to be more transparent about the use of the training data and potentially give copyright holders the option to block web crawlers and other software used to train LLMs from accessing their copyright data.

For more information and some historical context around copyright and generative AI, take a look at the related articles by Joanne Brook (Data scraping for AI development) and myself (Copyright and copywriting: AI challenges).

Implementation of the Online Safety Act

We covered the introduction of the Online Safety Act towards the end of 2023. Several regulatory provisions are coming into effect, following the publication in December of The Online Safety Act 2023 (Commencement No. 4) Regulations 2024 and related SIs which define thresholds for service providers. Some of the key deadlines are:

• 16 March 2025 – relevant providers must assess the risk of illegal harms on their services
• 17 March 2025 – illegal harms safety duties become enforceable by Ofcom
• 16 April 2025 – relevant providers must carry out a “children’s access assessment” to gauge whether its services are likely to be accessed by children
• April 2025 – Ofcom will publish its Protection of Children Codes and risk assessment guidance
• July 2025 – services which allow user-generated pornographic content must have “highly effective age assurance” (HEAA) systems in place (this was already a requirement for services which publish their own pornographic content, as of 17 January 2025).

Ofcom has published a range of guidance and materials related to the Online Safety Act, along with digital toolkits. It’s also worth noting a recent government commissioned review which suggests providing Ofcom with greater powers for banning certain types of adult content.

Fake online reviews

Back in 2021, the Competition and Markets Authority (CMA) launched an investigation into Google and Amazon over concerns it may have been breaching consumer law by failing to take sufficient action to protect people from fake reviews. In response to this investigation (finally), Google has signed undertakings which require the company to sanction UK businesses that have boosted their star ratings with fake reviews, as well as individuals who have written false reviews of businesses.

Meanwhile, the High Court in Ireland has ruled that an individual who posted a range of malicious accusations against a plumbing company on Google reviews and Trustpilot, was guilty of defamation and must pay the company €40,000 in compensation.

Conveyancing modernisation

The Ministry of Housing, Communities and Local Government has announced fresh plans to reform the conveyancing process, in conjunction with a “collection of industry and government experts committed to digitalising the home buying and selling process” known as the “Digital Property Market Steering Group”. The press release notes that the HM Land Registry will “build on its work in digitising property information and lead 10-month pilots with a number of councils to identify the best approach to opening up more of their data and making it digital, whilst the government pushes ahead with plans for digital identity verification services including in the property sector.”

This isn’t the first time that attempts have been made to streamline the UK home buying process through the application of technology, as wryly noted by The Law Society Gazette.

Removal of encryption by Apple

Following pressure from the UK government to be given access to user data, Apple has pulled its end-to-end encryption tool – Advanced Data Protection (ADP) – from the UK. This essentially means that the data of Apple users will be less secure in general; as well as security services being able to view data, it will also be more open to malicious hackers.

Commenting, Professor Alan Woodward, a cybersecurity expert at Surrey University, said the move was a “very disappointing development” and claimed that: “All the UK government has achieved is to weaken online security and privacy for UK based users.” The move has also created waves across the pond, with the CIA warning that the UK’s request to view international data would constitute a “clear and egregious violation” of American’s privacy and civil liberties.

Legal AI roundup

Aside from the AI Action Plan and consultation mentioned above, further recent developments related to AI and law include:

Global AI summit

In contast to all the fanfare of the first global AI summit held in Blethchley Park back in 2023, the 2025 summit in Paris was a bit of a damp squib, with the UK and US refusing to sign a proposed global declaration on AI. However, an International AI Safety Report, commissioned as a result of the Bletchley Park summit – which is the “culmination of work by 100 AI experts to advance a shared international understanding of the risks of advanced Artificial Intelligence” – has now been published.

Law firm bans AI

Hill Dickinson has blocked access to several generative AI tools after it “noticed a significant increase in usage of, and uploading of files to, such tools”.

Deepseek

The Chinese LLM, which was supposedly developed for a fraction of the cost of its rivals such as ChatGPT, took the world by storm when it was released in January 2025. But data protection regulators globally have been raising concerns regarding potential Chinese state control, and apparently the ICO is conducting an investigation.

Copyright of AI output

According to a report by the US Copyright Office, creative works which are produced with the assistance of AI tools could enjoy copyright protection.

Agentic AI

2025 has been touted as the year of “agentic AI” which essentially describes the ability of AI products to take actions on behalf of users which align with their goals, such as making purchases or sending emails. Hot on the heels of its own version of generative AI, LexisNexis has unveiled (currently just for US users) a personalised agentic AI assistant known as LexisNexis Protégé.

European news

A couple of recent EU internet law developments include:

Latest Schrems campaigns

Well known figure in the European data protection sphere, Max Schrems, is continuing his fight to secure data privacy for citizens against the clutches of big tech algorithms. One of his recent cases has resulted in a ruling from the Court of Justice of the European Union which chastised Meta for not adhering to data minimisation principles. Meanwhile, his data protection organisation called “noyb” (None of Your Business), has filed several complaints against companies transferring the data of European citizens to China.

Closure of online dispute resolution platform

The European Online Dispute Resolution Platform, which provided a system for consumers and traders to reach an out-of-court resolution of disputes arising from online sales or service contracts, is being closed on 20 March 2025. The reason for the closure is apparently due to very low take up.

Alex Heshmaty is technology editor for the Newsletter. He runs Legal Words, a legal copywriting agency based in the Silicon Gorge. Email alex@legalwords.co.uk.

Photo by Simon Berger on Pexels.