The UK government has officially abandoned its plan to allow AI companies to train on copyrighted works without explicit permission. The proposal, which would have introduced an opt-out model requiring rights holders to actively block access to their work, was announced at the end of 2024 and immediately sparked a fierce backlash.

Artists including Sir Elton John and Dua Lipa publicly opposed the policy, joining publishers, record labels, and media groups who argued it would hollow out copyright protections and undermine the UK's £146 billion creative sector.

Technology Secretary Liz Kendall confirmed the reversal on Wednesday, saying the government had "listened" and that the opt-out approach was "overwhelmingly rejected." She stated plainly: "The government no longer has a preferred option."

The climbdown aligns the government with the position of the House of Lords Communications and Digital Committee, which this week warned there was "no sound basis" for weakening copyright through opt-out mechanisms and called instead for a licensing-first approach. Peers also pushed for stronger transparency requirements so creators can see how their content is used in AI training.

The move leaves UK AI policy in an uncertain state. Ministers say they won't reform copyright law "until we are confident that they will meet our objectives," but critics in the tech sector warn that the delay risks the UK falling behind international competitors already building clear AI regulatory frameworks. Tech UK's deputy chief executive Anthony Walker said the UK "cannot afford for this to remain unresolved."

The UK AI industry and its creative sector remain at an impasse — but for now, existing copyright law holds: AI companies cannot use protected works for training without permission.