The UK’s creative industries are a global success story – driving employment, developing skills, and contributing significantly to GDP, all while delivering world-class entertainment to audiences at home and abroad. From film and television to music and journalism, these sectors enrich our cultural life and support millions of jobs. With the potential to generate an additional £10 billion annually by 2033 (Source: Sky/Public First), their economic importance is only growing. Yet AI presents both exciting opportunities and serious challenges to the future of creative work. 

Last year, fans of Studio Ghibli were stunned to see AI-generated clips circulating online that looked like they had been lifted straight from Spirited Away or My Neighbour Totoro. These were not lost treasures from Ghibli’s vaults – they were new imitations, generated by AI systems seemingly trained on the studio’s copyrighted films without its knowledge or consent. Similar stories are now commonplace across the creative industries. For large organisations, unlicensed use undermines hard-earned investment. For small creators, it threatens their survival. For artists, misappropriation of their voice or likeness puts their livelihood at risk. Left unchecked, this hollows out the creative economy – one of the key growth-driving sectors identified by the Government – and erodes trust in the industries that produce films, music, sports, and news. 

When we first published this piece in November 2025, the central debate was whether frontier AI companies should pay for the training data their models consume. The Government had floated an “opt-out” model – creating a default rule that AI developers could use copyrighted material to train AI models unless creators took active steps to reserve their rights. The creative industries pushed back, and rightly so. 

In March 2026, the Government published its report on copyright and AI, required under the Data (Use and Access) Act 2025. The opt-out proposal has been abandoned. There is no preferred legislative option. The Government is monitoring how the US and EU resolve these issues before committing to reform. 

We understand the caution – poorly-designed legislation can cause more harm than good. But the window to shape a fair, functioning rights marketplace, before informal norms solidify and market power concentrates further, is narrow. The House of Lords Communications and Digital Committee put it plainly in its March 2026 report: the UK’s creative industries face a clear and present danger from generative AI trained on copyright-protected works without authorisation or remuneration. The Committee called for a licensing-first approach backed by statutory transparency obligations, and firmly rejected the proposed opt-out model. Sky agrees.

The argument is no longer primarily about payment for training data used to build frontier models that already exist. The outcome of ongoing litigation will determine how existing legal rights apply. For the UK, however, the real opportunity is different – and arguably more important: creating a framework for the future licensing and deployment of AI. 

First, ongoing use: AI systems continuously ingest, retrieve and reproduce creative content in their outputs – from news summaries to video generation – every day, at scale. Second, market structure: will a functioning licensing marketplace emerge, or will AI companies access content for free while the creative sector loses revenue? That outcome is not pre-determined – it depends on the rules set now. Third, transparency: rights-holders are inhibited from enforcing rights where the functioning of models is opaque. Without mandatory disclosure of training data and retrieval sources, a licensing market cannot function. 

The question is no longer whether AI will use creative content. It is whether creators will be fairly paid when it does.

The most significant development since our last paper is the Creative Content Exchange – a marketplace for licensing creative and cultural assets for AI training and other uses. The pilot launched in December 2025, hosted by the Natural History Museum with UK Research and Innovation (UKRI) and Department for Culture, Media and Sport (DCMS), with a fuller platform expected by summer 2026. 

The CCE matters because it is the most credible mechanism for turning fair compensation into reality at scale. Rights-holders – including independent creators who cannot afford to negotiate bilateral deals with major AI companies – get a route to market. AI developers get verified, high-quality, legally-clean content. Done well, it is a new engine of economic value. 

The CCE’s potential scales significantly when paired with transparency obligations on AI developers. Rights-holders need visibility of how their content is being used to participate confidently in licensing markets. Extending the CCE to commercial creative industries – broadcasters, publishers, music rights-holders and sports content owners – and pairing it with disclosure requirements for AI developers would unlock that potential. Sky stands ready to engage.

The UK has an opportunity to draw on approaches being developed in the US and EU while charting its own course. The decisions made now will shape the UK’s position in one of the most consequential regulatory debates of the decade – and the UK’s creative and AI sectors both stand to benefit from getting it right. 

The EU has moved significantly further than the UK in establishing a framework for AI transparency. Providers of general-purpose AI models are required to publish a detailed summary of the content used to train their models, backed by significant penalties for non-compliance. While the EU's opt-out approach to commercial text-and-data-mining has its shortcomings, greater transparency will create real pressure on AI developers to engage with licensing markets. The UK should build on those lessons with a framework that combines transparency, legal certainty and fair compensation. 

In the United States, AI copyright continues to be litigated through the fair use doctrine, with cases multiplying rapidly. The legal uncertainty is generating significant risk for AI developers – precisely the kind of pressure that makes licensing frameworks attractive, if the UK makes one available. The New York Times’ lawsuit against OpenAI – now approaching its third year, with no resolution in sight – illustrates the significant legal risk AI developers face when building models without content licences. 

There is also a specific gap the UK should address. The UK’s copyright regime has limited reach over AI systems trained abroad, but systems deployed in the UK, generating outputs consumed by UK audiences – are a different matter. Clarifying that AI systems generating outputs in the UK are subject to UK copyright obligations, regardless of where the model was trained, would strengthen the overall framework.

The implications of AI for journalism go beyond creative and economic rights – they touch on trust, democracy and the integrity of public discourse. News organisations like Sky News invest in rigorous, impartial journalism, yet AI systems increasingly summarise and synthesise journalistic content, often with limited attribution, without commercial arrangements with publishers and no guarantee of accuracy. AI models can misattribute facts, hallucinate stories, and cite reputable sources for content they never published. 

Amid these risks lies a real opportunity. AI and semantic search could reset how the internet surfaces trusted information – rewarding credibility over volume. Publishers must help design the next generation of AI interfaces, especially around video content, ensuring quality journalism is discoverable, attributed and monetised. Any rights marketplace should treat journalism as a priority, not an afterthought. 

Sky News is playing an active role in building that marketplace, earlier this year co-founding SPUR — the Standards for Publisher Usage Rights coalition — alongside the BBC, Financial Times, The Guardian and Telegraph Media Group. With expanded international membership from Europe and the US, and publishing a draft content standard for public comment, SPUR is developing the technical infrastructure to track how AI systems use journalistic content — so that licensing and fair compensation can follow.

As one of the UK’s leading owners of entertainment, news, and sports content, Sky believes the stakes could not be higher. The Government estimates that the UK’s creative industries contribute more than £125 billion annually to the economy and provide over 2 million jobs. They define Britain’s cultural influence on the global stage. AI can and should support this sector – but only on a foundation of respect for intellectual property. 

Sky’s five principles for the UK’s AI copyright framework remain unchanged. What has changed is the urgency. 

Permission – Creators must retain control over their work. Explicit permission – not opt-out – should be the default for use of copyrighted material in AI systems trained or deployed in the UK. The abandonment of the ”opt-out” model is the right outcome and should not be revisited. 

Transparency – Mandatory disclosure of training data sources is the precondition for everything else. The UK should implement transparency obligations at least equivalent to the EU AI Act, applying to models trained or deployed in the UK regardless of where the developer is headquartered. 

Fair Compensation – Creators should be paid fairly for the use of their work. The CCE is the most promising mechanism for making this real at scale – but fair compensation requires a market that functions, and markets require both transparency and enforcement. 

Attribution – Where AI outputs draw on creative works – including news content – users should be able to trace the source. The Government’s work on AI output labelling should move to binding standards. Technical standards such as C2PA, which verify the origin and authenticity of content, should be explored and adopted. 

Enforcement – Strong penalties, audit rights, and designated regulatory oversight are essential – including over AI outputs generated in the UK from models trained abroad. Rights that cannot be enforced are not rights.

The opportunity is clearest when you consider what a well-designed framework delivers for everyone. The current uncertainty – in which AI developers face legal risk and rights-holders face unclear routes to compensation – serves neither side well. A clear framework creates the confidence for markets to scale. Acting now is in the interests of the whole AI and creative ecosystem. 

A licence-first framework accelerates growth by reducing uncertainty and building trust. For AI developers, it provides legal clarity – a transparent licensing regime gives them a clear path to the content they need, reducing expensive litigation risk. For the AI ecosystem, it means access to higher-quality data: licensed content from trusted newsrooms, broadcasters, film studios and sports rights-holders, rather than scraped material. For the broader economy, it creates a new marketplace in which both large organisations and independent creators can monetise their works at scale. Nascent licensing markets are already emerging between AI developers and major publishers. A clear framework would accelerate and formalise that trend. Above all, it builds public trust: a system that rewards creativity, ensures transparency, and lets consumers trace sources will drive faster adoption of AI across society. 

Taken together, this positions the UK not as a restrictive environment but as a global destination for AI investment and deployment – one that combines innovation with fairness.

With its forthcoming AI Bill, the UK has a genuine opportunity to set a global standard. There are also practical steps the Government can take now, ahead of primary legislation, to build the foundations:

• Make the Creative Content Exchange succeed: commit to expanding it to include commercial creative industries; provide clarity on governance and commercial models; and design the pilot to scale. 

• Build transparency into the foundations: disclosure of training data sources for models deployed in the UK – achievable through secondary legislation or regulatory guidance – is the foundation on which a functioning licensing market depends. Early action here also aligns the UK with EU transparency requirements already in force.

• Align AI output labelling with international standards: aligning with the EU’s Article 50 obligations – which requires AI-generated content to be clearly marked as such - supports consistency for businesses operating across both markets and gives consumers confidence in what they are seeing.

• Extend copyright protections to AI outputs in the UK: clarifying that AI systems generating outputs in the UK are subject to UK copyright obligations – regardless of where the underlying model was trained – would strengthen the overall framework.

• Enshrine a licence-first framework in law: making explicit permission the default, with safe harbour for developers who participate in good faith in licensed marketplaces.

• Support smaller creators: ensure that the benefits of any licensing framework reach independent creators and smaller rights-holders, not just large organisations with the resources to negotiate bilateral deals.

• Establish a dedicated regulator with real powers: with audit rights, enforcement powers, and the ability to act against AI models trained or deployed in the UK – designed now so it is ready when primary legislation arrives.

• Drive international cooperation: engage actively with the EU and US to avoid regulatory fragmentation; common transparency standards in particular are achievable without waiting for comprehensive legislative alignment.

The UK is well placed to lead. The opt-out model has been set aside. The Creative Content Exchange is taking shape. International frameworks are crystallising. The moment to shape the global settlement – by demonstrating that a fair, functioning rights marketplace and a thriving AI sector are not in conflict – is now. 

Sky’s ambition is straightforward. A functioning marketplace in which creators participate fairly in the AI economy – underpinned by transparency that makes licensing and enforcement possible, and that gives AI developers the legal clarity and quality content they need to innovate. 

The countries that succeed in AI will not be those that weaken intellectual property, but those that create trusted markets in which innovation and creativity reinforce one another. A creative economy that thrives, an AI sector with the content and legal certainty it needs to innovate, and a global reputation for getting this right. That is a prize worth moving quickly for.