The UK government published its position on copyright and AI on 18 March 2026. After months of deliberation, industry consultation, and anxious waiting from both rights holders and AI developers, the verdict is: no reforms — at least not yet.

What the Position Says

The government has declined to introduce reforms to copyright law. It is not granting AI companies a blanket exception to train on copyrighted material. Instead, it is proposing a broad copyright exception with an opt-out mechanism — meaning rights holders would need to actively opt out if they didn’t want their content used.

This is being framed as a “flexible” approach. Critics would frame it as kicking the problem down the road.

What This Actually Means

The opt-out model sounds elegant in theory. In practice, it places the burden entirely on rights holders — photographers, journalists, publishers, artists — to proactively prevent their work from being consumed by AI training pipelines. Most of them don’t have the technical resources or legal knowledge to do this effectively.

AI companies, meanwhile, get to proceed with training runs unless individual rights holders intervene. The asymmetry is structural, not accidental.

The government says it “must take the time needed to get this right.” That’s a reasonable position in the abstract. But “taking time” while the industry moves at full speed is its own kind of decision — it favours the party with momentum.

The Industry Context

The UK has been trying to position itself as an AI-friendly jurisdiction. Sticking with the status quo on copyright — rather than following the EU’s more prescriptive AI Act approach — is clearly intentional. The government wants UK AI companies to have access to the same training data pools as their US competitors.

Whether that gamble pays off depends on whether you think copyright law is a meaningful constraint on AI capability, or whether capability is primarily bottlenecked by compute and data scale anyway.

The Unresolved Questions

  • Who pays for the opt-out infrastructure? Rights holders need systems to register and enforce opt-outs. The government hasn’t addressed who builds or funds these.
  • What counts as “training”? The position paper is vague on whether inference-time use of copyrighted material counts the same as training-time ingestion.
  • International coordination. The UK is making domestic policy while the EU, US, and Canada are all working on their own frameworks. A UK-specific opt-out system doesn’t solve the cross-border training problem.

The Take

The UK government chose the path of least resistance: keep the existing framework, add a mechanism that sounds empowering but places the burden on the wrong side, and call it “flexibility.”

Whether this attracts AI investment or invites legal chaos remains to be seen. The honest assessment is that the government bought itself time without solving the underlying problem — and that time may run out faster than it expects.

Sources: Osborne Clarke, Expert Zoom