UK Kills AI Copyright Opt-Out, Pivots to Licensing and Deepfakes
The UK government has officially dropped its previously preferred proposal to allow AI developers to train on copyrighted works by default with an opt-out for rightsholders, after strong opposition from creative industry stakeholders. The government will instead pursue further work on digital replicas and deepfakes, labelling of AI-generated content, creator transparency and control mechanisms, and support for licensing approaches—alongside potential changes to UK protections for computer-generated works.
Key Takeaways
- The UK no longer has a “preferred option” for copyright-and-AI, effectively shelving the opt-out training proposal.
- Next policy focus areas: digital replicas/deepfakes (summer consultation) and best-practice labelling of AI-generated content (taskforce with interim report in autumn).
- Government will review creator control and “input transparency” mechanisms—standards, technical solutions, and enforcement gaps.
- A working group will explore how government might help independent/smaller creatives license content for AI training.
- A notable additional signal: potential removal of UK copyright protection for computer-generated works, aligning more with US-style human-authorship debates.
Why It Matters
For streaming, this is a clear “opt-out is dead; licensing is the new interface” moment. If UK policy coalesces around licensing plus transparency, AI video features (localization, metadata, trailers, synthetic dubbing, personalization) become less about model capability and more about rights provenance, audit trails, and deal infrastructure. Expect pressure on platforms and vendors to prove training inputs, label synthetic outputs, and manage consent for likeness—especially for talent. The wildcard is computer-generated works: changing that rule could reshape ownership of AI-assisted assets, complicating monetization and catalog strategy.
Read full article at broadcastnow.co.uk