United Kingdom

[GB] Report on copyright and artificial intelligence presented to UK Parliament

IRIS 2026-5:1/8

Julian Wilkins

Wordley Partnership

On 18 March 2026 the government published its Report on Copyright and Artificial Intelligence (the Report), which follows a consultation that ran from 17 December 2024 to 25 February 2025. Under section 136 of the Data (Use and Access) Act 2025 (the Act), the Secretary of State for DSIT and DCMS must, before the end of the period of nine months starting from the day on which the Act was passed (19 June 2025), prepare and publish a report on the use of copyright works in the development of Artificial Intelligence (AI) systems, and present this report to parliament.

The consultation received 11 520 responses from a broad range of stakeholders, including creators, rights holders, and AI developers, giving their views as to how copyright law should be adapted to take account of AI, including machine learning.

The Report does not set out reforms to the law but highlights the overall findings of the consultation. The Report concludes that the government is limited about introducing new laws without further investigation and that international approaches to AI and the fast-developing AI technology would make it premature to introduce legislation at this time. With this in mind, the government will continue to monitor issues concerning AI in the context of copyright law particularly from the perspectives of creative industries. 

Within the Report there was no indication that an exception to copyright law to enable text and data mining (TDM) should be introduced. The Report indicates that the government believes that IP rights holders should be “fairly remunerated” for the value they add to the AI supply chain but, for now, there will be no new copyright exception for AI training. In the consultation, the government had preferred a broad TDM exception with IP rights holders being able to opt out. While this was opposed by the creative industries, some responses to the consultation also suggested implementing a broad TDM exception. For example, by including a “focused exception” to copyright that would support commercial science and research (an extension of the existing non-commercial research exception), or a public interest exception that would permit AI tools to use copyright content for the purposes of detecting harm. However, the Report stated that any exception would only apply to material that had been lawfully acquired and accessed (i.e. not pirated). The Report finally suggested that if such an exception were to be made in law it could include a statutory remuneration model for rights holders.

The government’s current approach is to develop best practice around the transparency of training inputs, which it sees as a prerequisite for both rights enforcement and a functioning licensing market. The government aims to test commercial models for licensing as part of the “Creative Content Exchange” announced last year, and an operational pilot platform should be ready by summer 2026.

The government further indicated its preferred approach of removing copyright protection for wholly computer-generated works with no human author, while retaining protection for AI-assisted works where a human has contributed creatively. It says that this is consistent with the principle that copyright “should incentivise and protect human creativity”. This approach was widely supported by most of the contributors to the Report.

The Report also addressed the issue of digital replicas, namely AI-generated imitations of a person’s voice or likeness, identifying the inadequacy of existing copyright and performers’ rights provisions. The government intends to explore options to combat the risks of impersonation for both artists and the public who may be misled by AI-generated imitations, including, for example, the option of creating a new “personality right”. The Report acknowledged that more well-known artists may be able to protect their voice or likeness in the UK via the tort of passing off or registering trademarks, but for lesser-known artists and the general public, this will be insufficient due to a lack of public awareness and the absence of commonly identifiable characteristics compared to those of a well-known performer.

Other issues considered in the Report included greater transparency as to how AI developers train their learning models, including the content and data they use. The Report advocates promoting clarity and enforcement for rights holders, without disproportionate effects on AI development or deployment in the UK. The government will work with the industry and experts to develop best practice and inform any future potential legislation.

Labelling content to show whether AI was used may help protect against disinformation and harmful deepfakes. Currently, there are no obligations in the UK for AI-generated content to be labelled, but many services already include labelling technology, and several countries have introduced labelling rules. The government would continue to monitor international best practice to help develop appropriate rules.

Finally, the Report identified the need for the effective enforcement of rights, ensuring that any new rights are accessible to all rights holders and proportionate. It also recommended working with various parties, including law enforcement and the judiciary, to help ensure the UK enforcement framework remained fit for purpose.