The UK government has stepped back from a previously favoured opt-out model for the use of copyright works in training artificial intelligence, leaving the current legal framework unchanged while it gathers more evidence and watches how the courts and other jurisdictions develop the law. According to the Department for Science, Innovation and Technology’s report published on 18 March 2026, the consultation exposed deep divisions between creative rights holders, who want consent and payment at the centre of any regime, and AI developers, who argue that compulsory licensing could slow innovation and weaken Britain’s position in the sector.

The report does not resolve the core question of whether training an AI system on protected works amounts to infringement, and that uncertainty continues to be tested in litigation. Rather than rushing to legislate, ministers said they would continue evidence gathering, technical work and monitoring of international developments, a shift that legal commentators have described as a reset in policy direction.

Transparency emerged as one of the few areas where there was broad convergence. The consultation pointed to the need for greater, but proportionate, clarity over training data, the provenance of outputs and accountability mechanisms, with the government noting that these questions overlap with data protection, consumer protection and wider AI governance rules.

On computer-generated works, the government said most respondents did not support copyright protection for works created solely by AI, while leaving open the possibility that material produced with meaningful human involvement should continue to qualify. The report also suggests that labelling of AI-generated content is gaining acceptance, although ministers stopped short of proposing new mandatory rules and said they would instead monitor best practice and international standards.

The issue of digital replicas and deepfakes was treated more cautiously. The government acknowledged growing concern about the unauthorised use and commercialisation of a person’s likeness in a system that lacks a single, dedicated image-rights regime, but said it had no immediate plan to introduce personality rights. Instead, it will gather views on whether the current patchwork of laws remains fit for purpose. For rights holders, the practical message is to keep tightening access controls and contractual protections; for AI developers, the report is a reminder to strengthen governance, record-keeping and risk controls while the law continues to evolve.

Source Reference Map

Inspired by headline at: [1]

Sources by paragraph:

Source: Noah Wire Services