Normal View Dyslexic View
Commentary
A capitalist contest: the AI industry v. the creative industries
Abstract
This paper examines whether artificial intelligence industry developers of large language models should be permitted to use copyrighted works to train their models without permission and compensation to creative industries rightsholders. This is examined in the UK context by contrasting a dominant social imaginary that prioritises market driven-growth of generative artificial intelligence applications that require text and data mining, and an alternative imaginary emphasising equity and non-market values. Policy proposals, including licensing, are discussed. It is argued that current debates privilege the interests of Big Tech in exploiting online data for profit, neglecting policies that could help to ensure that technology innovation and creative labour both contribute to the public good.
Keywords
artificial intelligencecopyrighttext and data miningcreative industriessocial imaginaryinnovationpublic goodBig TechIntroduction
Large Language Models (LLMs) developed by the artificial intelligence (AI) industry are trained on online works produced by the creative industries, typically without prior permission, recognition or compensation to rightsholders. AI companies operating in commercial markets deem this practice essential if they are to build a commanding market share and achieve the profit margins they promise to their investors. Unhindered access to copyright-protected online information is an essential element of these companies’ business model, which they claim promotes both science and the ‘useful arts’ by providing AI tools and applications, including generative AI (GenAI), for the benefit of society. This is said to be facilitating a flourishing ‘democratisation’ of content creation that can help to avert an ‘algorithmic monoculture’ (Kleinberg & Raghavan 2021). If copyright law precludes these companies from scraping information from the Internet without seeking permission and compensating rightsholders, or the law is modified in a way that increases the costs of doing so, the AI industry argues that legislation either needs to be interpreted in their favour or it needs to be changed.
In this extended comment, I consider why this AI industry view is hotly contested by the creative industries, the values at stake as options for change are proposed and some of the broader implications for the future of our datafied world. The issues are examined in relation to policy debate in the United Kingdom (UK), which has focused on the right to access creative online works using text and data mining (TDM) to obtain data for input to LLMs. In the next section I use a ‘social imaginary’ concept to frame the contest between the AI industry and the creative industries. I follow this with a discussion of how values embraced within a dominant AI innovation and creative industries social imaginary yield aspirations that shape perspectives on the TDM controversy. Policy proposals and AI industry and creative industries responses are discussed next. In the conclusion, I reflect on how policy debate around TDM practice supresses an alternative social imaginary of AI innovation and creative labour, which values their non-market contributions and potential.
A social imaginary framing of AI industry versus creative industries conflict
A ‘social imaginary’ concept is helpful in framing contentious issues around AI industry practices and their impacts on the creative industries.1 A social imaginary refers to normative notions about how ‘things go on among people’. The concept can be applied to reveal the social, political and economic values that are taken as ‘common sense’ and which guide actors’ expectations and aspirations (Taylor 2004). In any period, dominant and subordinate social imaginaries are present and contests arise among adherents to different value sets. In the case of digital technology development, values embedded in a dominant social imaginary foster data-driven economic growth. In this imaginary, risks and harms attributable to digital innovation tend to be externalised and it is assumed that the interests and aspirations of societal actors can be ‘balanced’ through policy measures. When these interests are in contention, the expectation is that disputes will be resolved principally through the (neoliberal) capitalist market, enabled by non-restrictive governance rules (Mansell 2012). In the case of AI innovation, this imaginary helps to sustain a ‘seemingly common-sense yet hegemonic understanding of an “ethics” of AI/ML [machine learning]’. AI developments are seen as ‘world-historical forces of change-inevitable seismic shifts to which humans can only react’ and this view becomes normalised (Greene et al. 2019: 2122, 2127).2 As Nowotny (2021: 20) has observed, ‘when self-fulfilling prophecies begin to proliferate, we risk returning to a deterministic worldview in which the future appears as predetermined and hence closed’. In this social imaginary, efficiency gains arising from AI techniques of optimisation are highly valued, as is the project of reconstructing human behaviour through a ‘transcendent (“virtual”) order of mathematical formalism’ (Agre 1994: 107; and see McGuigan 2023). This imaginary embraces ‘vain pretensions … to understand mind as computation’ (Winograd 1990: 167; and see Mansell 2024). Similarly, within this imaginary, interest in the sustainability of creative and cultural work is positioned, as Schlesinger (2017: 73) writes, principally as ‘an overwhelmingly economic subject matter whose central trope is the “creative economy”’, and individual creativity is positioned as a source of wealth creation.
Framed by an alternative AI innovation and creative labour social imaginary that privileges non-market values, the negative impacts of AI innovation are prominent—such as challenges to fundamental rights and the need to secure privacy, freedom from surveillance and data-related discriminations, as are concerns about the sustainability of those who are not leaders in the AI industry or in the creative sphere (Crawford 2021; Mansell & Steinmueller 2022; Mejias & Couldry 2024; Zuboff 2019). In this typically subordinated imaginary, attention may be given to ethical or responsible AI development principles, but it is not presumed that policy will balance interests, especially not with respect to disadvantaged and marginalised populations. Universalising ethical principles are criticised for drawing on values asserted to be ‘common sense’ by those in the Global North (Mansell et al. 2025: Ch. 3). Those aspiring to global leadership in AI and creative industries markets in the Global North are criticised for being blind to an imposition of ‘the technocratization / instrumentalization of Coloniality / modernity’ (Quijano 2017: 365) through AI advancement.
The core of this alternative imaginary is a valorisation of equity and fairness. Redress for harms associated with market-oriented AI innovation is said to require an undoing of ‘the logics and politics of race and coloniality that continue to operate in technologies and imaginaries associated with AI in ways that exclude, delimit, and degrade other ways of knowing, living, and being that do not align with the hegemony of Western reason’ (Adams 2021: 190). In respect of creative labour, in this imaginary what is valued is, for example, the contributions of creative work to social inclusion, its educational value and its aesthetic value. In brief, it is the public value of cultural production and non-market-based logics of production that are privileged instead of the commodification of culture (Hesmondhalgh & Pratt 2005). Capitalist market logic is criticised for diminishing non-market, collective or public logics of production, whether these involve journalism or other facets of the creative arts (Pickard 2020). This critique is amplified in efforts to reimagine how the agency of cultural and creative workers can be sustained and decolonised in the Global South (Moyo 2024).
The next sections focus mainly on aspirations and contentions among actors whose expectations and aspirations are aligned with the dominant AI innovation and creative industries social imaginary to highlight how disputes around TDM practice in the UK play out in this context.
AI industry and creative industries aspirations
What is intended by ‘AI’ is often unclear. The focus in the policy debates about TDM is principally on AI industry practices relating to building LLMs and GenAI applications (a type of AI that generates text, images, audio, code, video and synthetic data). However, AI encompasses a complex value chain that is at the heart of what AI industry promoters claim to be a ‘global race for cultural and technological leadership [which] remains wide open’ (Mokander et al. 2025: 1). The aim of the AI industry companies is to generate profits by ‘industrialising’ or integrating LLMs into digital apps and services. The companies developing GenAI applications (for example, OpenAI or Anthropic) depend upon on Big Tech cloud infrastructures and investment. Mentions of AI and machine learning (ML) are present in the product pages of Amazon Web Services, Microsoft Azure and Google Cloud Platform, promoting applications of GenAI in sectors from government to education, healthcare, gaming, media, energy, manufacturing and more, all of which are sources of AI industry revenue (van der Vlist et al. 2024).
While revenues directly attributed to LLM development may not yield a profit, this technology is the core of the AI industry’s business, which involves a complex value chain with data positioned as the ‘key leverage in AI partnership power dynamics’ (Gupta et al. 2025: 63, Rieder 2022). Companies such as Apple, Microsoft, Alphabet, Amazon, Meta, Nvidia and Tesla, with a combined market capitalisation of some US$18 trillion in early 2025 (Krause 2025), want to ensure that their AI investments are governed in a way that does not jeopardise their prospects for growth. This is especially so when they face competition from developers of lower cost LLMs, for example, the Chinese Deepseek company’s AI model.
To succeed, and consistent with the dominant social imaginary, countries must harness the power of AI so that they can ‘set the technical, aesthetic and regulatory standards for others to follow’ (Mokander et al. 2025: 1). As a medium-sized country, the UK cannot risk being left behind, not the least because the AI industry in the UK is estimated to have generated £14 billion to 2024 and is projected to add £550 billion to GDP by 2035 (Walker & Verseckaite 2025).3 However, the rewards from AI innovation are far from certain and there are those who argue that hyped-up market valuations have created a tech bubble which could burst and destroy the leadership positions of today’s leading AI companies (Floridi 2024; Widder & Hicks 2024). The AI industry’s response is to self-promote their ethical behaviour and to advocate governance rules that do nothing to jeopardise AI investment (Krause 2025). In this context, the LLM developers want the ‘right to learn’ from the information (data) they find online without policy-imposed restraints.
The inevitability of a ramping up of marketised AI applications is positioned as requiring other industries, including the creative industries, to adjust. The creative industries in the UK contribute some £108 billion in annual gross value added to the economy (UK House of Commons 2024) with creative services global exports of US$1.1 trillion in 2020 and creative goods of US$524 billion (UK House of Lords 2023). LLM developer practices are seen as a threat to the creative industries because ‘the rapid growth of generative AI technologies threatens their earnings and future employment opportunities. This is not just an issue for one part of the industry: it about real lives and livelihoods, and the impact will be felt by the most vulnerable’ (UK House of Commons 2025: 79), with the expansion of the AI industry reported in the press as an ‘existential threat’ (Savage 2024).
The creative industries comprise large and small companies and independent creators, but smaller and independent actors (and some of the larger companies) have yet to become bound into the AI industry value chain and much creative labour is self-employed freelance work, which attracts poor remuneration.4 This is not solely attributable to the rise of the AI industry, but LLM and GenAI development practices are seen as exacerbating a squeeze on individual creators’ incomes and as threatening the creative industries’ sustainability (Kretschmer et al. 2025; UK House of Commons 2024). Similarly, in the United States (US), AI development is seen as likely to ‘obliterate the markets for much human creation’ and ‘threatens to dilute human creators’ royalties’ (Universal Music Group, cited in US Register of Copyrights 2025: 35, 37). Since passage of the Statute of Anne 1710 in the UK, copyright law has been intended to foster the creation of works by authors/creators, and, as Cornish (2002: 12) asks, ‘if we are not prepared to provide legal buttresses for the interest of the author, why are authors there at all?’ In the UK policy debate around TDM, most creative industries participants have called upon the Government to acknowledge their intellectual property rights and to ensure that the copyright regime is enforced to enable them to receive compensation from the AI industry for their labour.
There can be value misalignments within a given social imaginary. Thus, in the dominant social imaginary, the inevitability of a commercial market for AI tools and applications is widely accepted by the creative industries, even if it is not clear that the interests of smaller creative sector companies and independent creative industry producers will be fairly treated. There is historical precedent for concern about whether the large entertainment companies will treat all artists fairly when they enter TDM licensing deals. When music publishers and recording companies negotiated licensing deals, for example, with new entrant music streaming companies and video sharing platforms, Google was reported to have paid more than US$400 million to a single music label. Small creative producers in the music industry argued that they rarely benefitted (Noble 2025). Large rightsholders generally have the greatest resources to enforce their rights, and research shows that individuals on average achieve incomes from their creative endeavours that are less than for those with similar educational status (Watt 2009). Thus, in earlier rounds of technological advance, changes in copyright law have not always given independent artists the power to control and receive financial benefit from the use of their work. Instead, changes have often resulted in more power accruing to ‘megacorporate IP feudal lords’ (Drahos & Braithwaite 2002; Giblin & Doctorow 2022; Hesmondhalgh 2019).
It is also argued that Western intellectual property laws support the appropriation of indigenous people’s creations without fairly enabling originators to share in the proceeds. Tehranian (2025: 2, 10) argues in The Secret Life of Copyright, that ‘the wielding of intellectual property rights constitutes an inextricable part of the exercise of wealth and power … while the law may tell us that all copyrighted works are created equal, it turns out that some are more equal than others’. Creative artists have been locked into contractual arrangements and suffered a variety of forms of ‘artist abuse’ in the large company entertainment industry with its ambitions to increase stock valuations (Morris 2022). In the light of this history, the current debate about TDM is another round in the ‘copyright wars’ (Horten 2025).
The voices of those championing ‘technological accelerationism’ in line with the dominant AI innovation social imaginary command disproportionate attention, and legislative action has been found to favour their interests (Caballero & Monje 2024). This is despite the fact that creative industries’ associations do claim to represent all interests and to acknowledge that large and small creative businesses, and superstars versus ‘ordinary’ creators, are different entities with very different interests. In the next section, I turn to how the policy debate around copyright and LLM development in the UK has been framed.
Current TDM exception flashpoint
A temporary copy exception (section 28A of the Copyright, Designs and Patents Act) in the UK allows for making temporary copies of ‘no independent economic significance’ when works online are accessed, permitting LLM training on public data. A non-commercial TDM exception was introduced in 2014 without an opt-out option. This has been interpreted to allow AI developers to train their LLMs on copyrighted data if the purpose is ‘non-commercial’ (UK Government, 1988 as amended, S. 29A). Rightsholders, however, argue that they invest to create the information/data assets that are used by LLM developers to build their commercial products and services (Glenster et al. 2025).
UK Government and creative industries claims
The UK Government claims that it seeks a ‘balance’ of interests for the creative industries and the AI industry (UK Government 2025a). During a consultation in 2022, debate focused on whether copyright legislation should be modified to allow AI companies to engage in TDM for any purpose and without explicit permission (Intellectual Property Office 2022). At the time, a creative industries ‘opt-out’ approach was characterised as taking ‘insufficient account of the potential harm to the creative industries’ (UK House of Lords 2023: para. 34). The harms envisaged included non-consensual incorporation of artistic material (copyrighted works, audio-visual performances, etc.) into LLM training datasets without fair compensation, an absence of opportunities for redress if artists’ moral rights are violated, and harms to creative livelihoods. Creative industries representatives argued in this consultation round that they had been inadequately consulted (Miltner & Highfield 2024).
A new policy consultation was launched in 2024, again with a view to achieving a ‘balancing’ of interests. This time the Government indicated that it favoured a change in legislation that would create a TDM exception and a rights reservation (opt-out) process that would allow creators to seek remuneration when their outputs are used to train LLMs (Intellectual Property Office 2024), an approach that seemed to be in line with ambitions set out in an AI Opportunities Action Plan (UK Government 2025b). Other options were presented with a promise that no change to existing copyright law would be made without ‘robust transparency measures’ (Intellectual Property Office 2024: 19). The ‘opt-out’ option has been seen by some analysts as delegating responsibility to rightsholders to identify rights violations (Piasecki & Helberger 2025). Others have argued that no effective rights-reservation opt-out process is available and that relying on metadata labelling and automatic content recognition will not deliver transparency (Davies 2025; Scott 2025). Instead, many representatives of the creative industries have called for an ‘opt-in’ system in the hope of increasing leverage to agree licensing terms with the AI companies (Savage 2024).
Broadsheet press coverage of the TDM dispute in the UK has tended to favour the proposals of the creative industries. This is unsurprising since news publishing companies are impacted when GenAI bots automatically summarise news at the top of web pages for social media users, distracting their attention from links to content that generates revenues for publishers (and advertisers).5 Coverage by legal press blogs generally raised concerns, but did not take a firm view of the options, and specialist AI tech and computing media outlets, mainly supported AI industry positions (with exceptions, since some tech experts are developing ways to encourage transparency and rewards to creators). There was little sign of coverage of how any change in UK copyright legislation might impact on creative industries actors beyond the UK (whether in the European Union or the Global South). There were observations on the impacts of changes in the TDM exception on foreign-owned companies, mainly considering whether the Government has the power to influence corporate behaviour without jeopardising its ambitions for the UK’s AI sector.6
Large creative industry companies and their associations (also representing small independents) responded to the consultation by arguing that copyright in creative works should be respected and enforced with enhanced transparency and the use of licences (Booth 2024; Glenster et al. 2025). This was the position, for example, of the Creative Rights in AI Coalition and the Society of Audiovisual Artists (2025), which also argued that ‘failure to act’ will ‘unquestionably have a catastrophic and irreversible impact on all UK authors’ (Creamer 2025: np).7 Individual creators responded to the Government’s consultation in large numbers and a Creative Rights Alliance (representing some 500,000 creators) called for transparency, enforcement of copyright law and remuneration through licensing (CRA 2024). A ‘Make IT Fair’ campaign launched by the (News Media Association 2025) was supported by Sirs Elton John and Paul McCartney and many others.
Any social imaginary can accommodate some diversity of values. Some in the creative industries treat AI development as an opportunity for economic inclusion. In this case, GenAI based on LLMs is seen as offering general-purpose tools that can benefit historically disadvantaged groups. If the AI industry is forced to pay for content to train their models, this could jeopardise inclusivity and the production of AI-enabled art forms less likely to replicate social and political biases: for example, artworks produced by the African American community and other groups (Ferrara 2024; Jackson 2025; Noble 2025).
Before reaching a decision on the appropriate course of action in the TDM dispute, the Government introduced a new Data (Use and Access) Bill. As the Bill proceeded through parliamentary debate in early 2025, amendments were tabled to secure AI industry compliance with existing copyright law. In the House of Lords, Baroness Kidron proposed an amendment: ‘the Secretary of State must by regulations make provisions clarifying the steps the operators of web crawlers and general-purpose artificial intelligence (AI) models must take to comply with UK copyright law’.8 During debate she said:
There’s no industrial sector in the UK that government policy requires to give its property or labour to another sector—which is in direct competition with it—on a compulsory basis, in the name of balance … Silicon Valley has persuaded the government that it’s easier for them to redefine theft than make them pay for what they have stolen (Kidron, cited in McKiernan 2025).
This and similar amendments were not accepted, and the Bill became law in June 2025 (UK Government 2025c).
The new Data (Use and Access) Act requires an economic impact assessment of options proposed in the Government’s 2024 Copyright and AI Consultation paper (Intellectual Property Office 2024) and other options considered ‘appropriate’ by the Secretary of State. Impacts on copyright owners and those who develop or use AI systems in and outside the UK are to be assessed, as are technical measures and standards to control the use of copyrighted works, the disclosure of information about such use, the use of licences and approaches to enforcement (UK Government 2025c).9 One option proposed by Kretschmer et al. (2025) is to broaden the existing TDM exemption to encompass all pre-competitive research and development that occurs before LLMs are integrated into commercial products. At the point of market entry, transparency and licensing obligations would apply, eliminating the distinction between data use for non-commercial and commercial purposes. When research is ‘industrialised’ and marketised in AI applications, this option calls for revenue sharing with those whose works are contained in datasets used to train the LLMs, and these authors recognised that identifying when an LLM enters the market is difficult.
In summary, reactions to proposals for changes in TDM practice are not uniform. The UK House of Commons Culture, Media and Sport (CMS) Committee observed that ‘proceeding with an opt out regime stands to damage the UK’s reputation among inward investors’ (UK House of Commons 2025: 77). The AI industry has continued to push for the broadest possible TDM exception without prior authorisation (Jackson 2025; Willemyns 2025), and, by April 2025, Open AI and Google had explicitly rejected an ‘opt-out’ approach (Jackson 2025).
European Union approach to the TDM dilemma
The European Commission published its GPAI (General Purpose AI) Code of Practice for implementation of the AI Act in July 2025 (European Commission 2025; European Union 2024), having introduced a TDM ‘opt-out’ provision in its 2019 directive on copyright and related rights in the digital single market (European Union 2019). The Code places voluntary requirements on AI providers to ensure transparency and copyright compliance. It was swiftly criticised by a coalition of forty organisations representing some seventeen million authors, performers and other creative industries rightsholders. They argue that the Code fails to require AI developers to train LLMs in a way that does not infringe on copyright or to provide a path towards a transparent rights reservation system (CISAC 2025), that is, an ‘opt-in’ arrangement. CISAC argued that ‘meaningful, actional transparency on the content used for training and other purposes, and its disclosure, is not only feasible but can be achieved with little technical or financial effort’ (CISAC 2025: np). The GPAI has been seen by AI companies and some AI-using companies as going too far with the risk of ‘slowing down Europe’s development and deployment of AI’ and there are calls for delayed implementation to avoid stifling innovation (Wilkinson 2025).
AI industry lobbying strategies and the courts
The AI industry’s lobbyists invoke ‘diversity’ issues to defend their current TDM practice, arguing that ‘if we want AI models that don’t replicate existing social and political biases, we need to make it possible for new players to build them’. They emphasise the importance of ensuring that small and large AI companies can affordably access the data needed to continue to train their LLMs (McSherry 2023: np). Licensing requirements, they suggest, will lock in the dominant positions of companies that can afford licences. Lobbying on these issues by AI companies such as Anthropic and OpenAI and the Big Tech companies such as Meta has ramped up substantially in the US in the wake of court cases (Lippman 2025; Marshall 2024) with outcomes in some instances that favour the AI industry. In June 2025, a court ruled that Anthropics’ LLM had not infringed copyright law, despite training its model on books without authors’ permission (
US District Court
2025a). Meta was not found to be using authors’ works to train its models in a way that results in ‘market dilution’, with a court ruling that it was not guilty of infringing copyright (Montgomery 2025).10AI companies argue that their use of copyrighted works to train LLMs is legal under the ‘fair use’ provisions of the US Copyright Act 1976 (Cooke 2025; Landymore 2025). However, they also claim that datasets ‘we know to be pirated’ have been used to train LLMs. Meta argued in Kadrey v. Meta that the books it scrapes for LLM training are worthless because each book makes a tiny change in the performance of its LLMs and that it does not benefit monetarily directly (US District Court 2025b). The positions taken by these companies present a contradictory picture and they are informed by lobbying strategies targeting policymakers. For example, a Meta internal document indicated that ‘if there is media coverage suggesting we have used a dataset we know to be pirated, such as LibGen, this may undermine our negotiating position with regulators on these issues’ (Landymore 2025).
The AI companies are strongly motivated to maintain the position that they do not directly benefit from the data they scrape from the Internet. They argue that AI innovation is at risk without access to relatively low-cost or free access to data. If lawsuits favour the creative industries, they claim that this could jeopardise their future economic viability. As discussed above, the hype surrounding GenAI development is part of a larger project of building dependencies and control across the whole of the digital infrastructure. By ingesting data without regard to copyright, vital resources can be enclosed within the control of big (and some smaller) corporate value chains.
Licence agreements and changes in TDM practice
In parallel with policy consultations, creative industries’ professional associations are devising codes of practice for the international and UK music industries and their independent creative producers (Booth 2024) and Kretschmer et al. (2025: 5) observe that ‘we have already moved to a licensing economy’ rendering moot the need for a change in copyright legislation. Larger players such as News Corp, The Financial Times and Reuters are negotiating data use licence agreements with AI industry companies. Most of some eighty licences examined in 2025 were with Open AI and Perplexity, with Big Tech companies (for example, Amazon, Apple, Google, Meta and Microsoft) entering fewer agreements. These licences include various types of rightsholder compensation (credit to authors, lump sums, percentages of revenue or per item payments).11
When HarperCollins proposed an arrangement to seek consent from authors whose works are to be included in a licence agreement with an AI company with payment per title, the Society of Authors responded that the split in remuneration between HarperCollins and individual authors needed further consideration (SOA Policy Team 2024). Kretschmer et al. (2025: 6) emphasise that ‘the interests of creatives (authors, artists, performers) … should not be bundled with the owners of large back-catalogues of rights’ and existing evidence on licence provisions do not indicate how much revenue will flow to the originators of content (Thomas & Kretschmer 2025; see also Glenster et al. 2025, Kahn 2025). It seems unlikely that agreements between the AI industry and creative industries players will fully and fairly take account of the interests of all rightsholders, especially in the face of pressure from venture capitalists to achieve healthy AI industry financial returns.
Initiatives are underway, however, that may boost the prospects for fair remuneration. The UK’s Copyright Licensing Agency (CLA), for example, is working on behalf of authors, publishers, designers and artists, and with the PLS (Publishers’ Licensing Services) and with ALCS (Authors’ Licensing and Collecting Society), to devise a text licensing system with a planned launch towards the end of 2025. This is expected to assist creative actors who are not well positioned to negotiate licensing deals with the AI industry (Cormack 2025; see also TUC 2025). In addition, a ‘Licensed Model Certification’ has been introduced by the non-profit organisation, Fairly Trained. This sets out requirements to ensure that LLMs are not trained on copyrighted works without a licence (Fairly Trained 2024). Certification requires LLM developers to use training data based on contractual agreements with records kept of the data that is used. Fairly Trained has so far certified text and speech generation models and is supported by The Authors Guild, SAG-AFTRA (Screen Actors Guild–American Federation of Television and Radio Artists) and the American Society for Collective Rights Licensing.12 Examples of certified AI developers are Beatoven.ai, a fully licensed generative music model funded by institutional investors in India and the UK (Beatoven.ai 2025); ElevenLabs’ AI music generator, which claims it is only trained on songs when a licence has been negotiated with owners of music catalogues (ElevenLabs 2025); and KL3M which claims to be the first clean family of LLMs trained on legally permissible data use. It asserts that ‘we know where every word in our training data came from’ and that it does not violate copyright legislation (KL3M 2025).
If decisions by the courts and/or policymakers do mandate a TDM ‘opt-in’, licensing of data subject to copyright and transparency of revenue distribution, it is unclear whether LLM developers will comply. In other digital governance domains, implementation of rules is uneven. Data privacy protection notices, for example, lack clarity years after the introduction of privacy legislation. The UK’s Information Commissioners’ Office found that the scale of data sharing was ‘disproportionate, intrusive and unfair’ and that ‘individuals have no guarantees about the security of their personal data within the [platform] ecosystem’ (ICO 2019: 4, 23) despite legislative requirements. Implementation and compliance of digital governance legislation are variable in the European Union, notwithstanding fines and other sanctions (Mansell 2021). And, the European Union’s AI Act provides ample scope for interpretive ambiguity, for example, it is expected to curtail only ‘reasonably foreseeable misuse’ of AI systems (Mansell in press). Thus, even if there is move in the UK to strengthen accountability and the transparency of AI industry TDM practice, this cannot guarantee compliance.
Conclusion
Is it inevitable that companies with the most financial clout and lobbying power will win the right to exploit online data for LLM training as they please? Critics of the UK Government’s position claim the Government is ‘siding with big US tech companies’ (Scott 2025: np). The voices of those representing the creative industries are heard by parliamentarians, but it is unclear how the TDM controversy will be resolved. The challenges facing the creative industries, and especially smaller companies and individual creators, are emphasised by the creative industries’ representatives and in the press and by some policymakers. Nevertheless, the drive to secure the prospects of the UK AI industry is strongly influenced by the dominant AI innovation and creative industries social imaginary, which presents a ‘common sense’ vision of a thriving commercial AI sector balanced alongside a successful creative industries economic contribution.
Resolving the conflict over TDM practices in the UK policy debate is presented by some policymakers and the AI industry as requiring the accommodation of ‘common sense’ consistent with the dominant social imaginary—an imagined future in which there is an inevitable incursion of marketised AI tools and services into all sectors, including the creative industries. In this context, declining incomes of smaller and independent contributors to creative production are at risk of being treated as collateral damage because the priority is to boost the AI industry’s and the creative industries’ contributions to economic growth. In view of the threat that LLM training will move to other countries if new policy is introduced, it may be that UK policymakers’ preferred option will be to do nothing, leaving TDM practices to be worked out in the marketplace.
The dominant AI innovation and creative industries social imaginary is remarkably insensitive to the reality of winners and losers in struggles over the distribution of rewards to creative labour. The way AI marketised products produce bias and discrimination is absent from the narrow TDM policy debate, apart from lobbyists’ claims that applications will enable enhanced diversity. ‘Common sense’ values within the alternative social imaginary are subordinated in favour of promoting Big Tech’s exploitation of data for profit. The alternative social imaginary calls for an emphasis on the future potential of AI innovation and creative labour to contribute to the public good. It embeds ‘common sense’ values associated with collective and non-market strategies and practices as a means of resisting the exploitation of data for profit.
The aim of this extended comment has not been to predict the outcome of the UK Government’s deliberations on TDM practice. It has been to reflect instead on the implications of the TDM debate as revealed through the lens of contending AI industry and creative industries social imaginaries. The discussion has highlighted how the TDM debate is influenced by a dominant imaginary that takes a logic of marketised AI innovation as ‘common sense’ alongside a strong emphasis on the commercial market for creative works. This imaginary has been contrasted with the ‘common sense’ embedded in an alternative social imaginary that encourages attention to the ambitions and expectations of those who are harmed. These harms are due to an emphasis on sustaining the wider project of ‘technocratisation’ and ‘instrumentalisation’ through commercial data exploitation that disadvantages groups in the UK and beyond. Policymakers who attain increased awareness of the importance of collective interests in equity and justice, the values embedded in the alternative social imaginary, may be more likely to begin imagining a future that attends to non-market approaches. In this way, the benefits of AI to creative labour can be developed, even if these emerge only in parallel to the commercial aspirations of both industries.
Acknowledgements
With thanks to Sonia Livingstone for helpful comments and to two anonymous referees for their comments on an early much shorter draft which were very helpful in expanding and clarifying the argument. The views expressed here are mine, including any errors or omissions.
References
AdamsR. (2021), ‘Can artificial intelligence be decolonized? Interdisciplinary Science Reviews, 46(1–2): 176–197. https://doi.org/10.1080/03080188.2020.1840225
AgreP.E. (1994), ‘Surveillance and capture: two models of privacy’, The Information Society: An International Journal, 10(2): 101–127. https://doi.org/10.1080/01972243.1994.9960162
Beatoven.ai (2025), Beatoven.ai: royalty free AI music generator. Beatoven.ai. https://www.beatoven.ai/
BoothR. (2024), ‘UK arts and media reject plan to let AI firms use copyrighted material’, The Guardian. https://www.theguardian.com/technology/2024/dec/19/uk-arts-and-media-reject-plan-to-let-ai-firms-use-copyrighted-material
CaballeroF.S. & MonjeD.I. (2024), ‘Artificial intelligence and technological accelerationism: a critique of cybernetic ideology’, in
Pedro-CarananaJ., GomezR., CorriganT.F. & CaballeroF.S. (eds), Political Economy of Media and Communication: Methodological Approaches (Routledge), 328–365. https://doi.org/10.4324/9781003385516-26
CammaertsB. & MansellR. (2020), ‘Digital platform policy and regulation: toward a radical democratic turn’, International Journal of Communication, 14(2020): 135–154. https://ijoc.org/index.php/ijoc/article/view/11182/2901
ChapekisA. & LiebA. (2025), Google Users are Less Likely to Click on Links When an AI Summary Appears in the Results (Pew Research Center). https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
CISAC (2025), Joint Statement by a Coalition of Authors, Performers and Other Rightsholders Active Across the EU’s Cultural and Creative Sectors Regarding the Third Draft of the EU AI Act’s GPAI Code of Practice (International Confederation of Societies of Authors and Composers). https://www.cisac.org/Newsroom/articles/joint-statement-coalition-authors-performers-and-other-rightsholders-active
CookeC. (2025), ‘Explainer: Donald Trump’s AI action plan’, TW Weekly. https://thisweekculture.com/article/explainer-donald-trumps-ai-action-plan/
Copyright, Designs and Patents Act (1988), Public General Acts. 1988, c.48. https://www.legislation.gov.uk/ukpga/1988/48/contents
CormackM. (2025), CLA Announces Development of Generative AI Training Licence (Copyright Licensing Agency). https://cla.co.uk/development-of-cla-generative-ai-licence/
CornishW. (2002), ‘The author as risk-sharer’, Columbia Journal of Law and the Arts, 26: 1–16. https://heinonline-org.lse.idm.oclc.org/HOL/Page?public=true&handle=hein.journals/cjla26&div=7&start_page=1&collection=journals&set_as_cursor=0&men_tab=srchresults
CRA (2024), The CRA Calls on Government to Protect Human Creativity in the Age of Generative AI (Creators’ Rights Alliance). https://www.creatorsrightsalliance.org/ai-and-creative-work
CrawfordK. (2021), The Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence (New Haven, Yale University Press). https://doi.org/10.12987/9780300252392
CreamerE. (2025), ‘Authors call for UK Government to hold Meta accountable for copyright infringement’, The Guardian. https://www.theguardian.com/books/2025/mar/31/authors-call-for-uk-government-to-hold-meta-accountable-for-copyright-infringement
Data (Use and Access) Act (2025), Public General Acts. 2025, c.18. https://www.legislation.gov.uk/ukpga/2025/18
DaviesC.W. (2025), AI Copyright Regime Steers Away From Requiring Licences in All Cases (Pinsent Masons). https://www.pinsentmasons.com/out-law/news/ai-copyright-regime-steers-requiring-licences-all-cases
DrahosP. & BraithwaiteJ. (2002), Information Feudalism: Who Owns the Knowledge Economy (Routledge).
ElevenLabs (2025), Webpage. https://elevenlabs.io/
European Commission (2025), Code of Practice for The General-Purpose AI Code of Practice (Copyright Chapter). https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai
European Union (2019), Directive (EU) 2019/790 on Copyright and Related Rights in the Digital Single Market. European Union, OJ L 130/92. https://eur-lex.europa.eu/eli/dir/2019/790/oj/eng
European Union (2024), Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 Laying Down Harmonised Rules on Artificial Intelligence … (Artificial Intelligence Act). European Commission, 12 July. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L_202401689
Fairly Trained (2024), Fairly Trained Licensed Model Certification (Fairly Trained). https://www.fairlytrained.org/certifications
FerraraE. (2024), ‘Fairness and bias in artificial intelligence: a brief survey of sources, impacts, and mitigation strategies’, Sci, 6(3): 1–15. https://doi.org/10.3390/sci6010003
FloridiL. (2024), ‘Why the AI hype is another tech bubble’, Philosophy & Technology, 37(128): 1–13. https://doi.org/10.1007/s13347-024-00817-w
GiblinR. & DoctorowC. (2022), Chokepoint Capitalism: How Big Tech and Big Content Captured Creative Labor Markets and How We’ll Win Them Back (Boston, MA, Beacon Press).
GlensterA.K., HamptonL., NeffG. & LacyT. (2025), Policy Brief: AI, Copyright, and Productivity in the Creative Industries (University of Cambridge, ai@cam). https://www.mctd.ac.uk/policy-brief-ai-copyright-productivity-uk-creative-industries/
GreeneD., HoffmannA.L. & StarkL. (2019), ‘Better, nicer, clearer, fairer: a critical assessment of the movement for ethical artificial intelligence and machine learning’, in Hawaii International Conference on System Sciences 2019 (HICSS-52). 2122–2131. https://aisel.aisnet.org/hicss-52/dsm/critical_and_ethical_studies/2; https://doi.org/10.24251/HICSS.2019.258
GuptaN., UrmetzerF. & AnsariS. (2025), ‘Big-tech strategic partnerships in artificial intelligence’, International Journal of Business and Management, 20(3): 57–72. https://doi.org/10.5539/ijbm.v20n3p57
HesmondhalghD. (2019), The Cultural Industries, 4th edn (Sage).
HesmondhalghD. & PrattA.C. (2005), ‘Cultural industries and cultural policy’, International Journal of Cultural Policy, 11(1): 1–13. https://doi.org/10.1080/10286630500067598
HortenM. (2025), ‘Copyright wars 3.0: the AI challenge’, IPTegrity.com. https://iptegrity.com/ai/copyright-wars-3-0-the-ai-challenge
ICO (Information Commissioner’s Office) (2019), Update Report into Adtech and Real Time Bidding (London, Information Commissioner’s Office). https://ico.org.uk/media2/migrated/2615156/adtech-real-time-bidding-report-201906-dl191220.pdf
Intellectual Property Office (2022), Consultation Outcome: Artificial Intelligence and Intellectual Property: Copyright and Patents (Intellectual Property Office, June). https://www.gov.uk/government/consultations/artificial-intelligence-and-ip-copyright-and-patents/outcome/artificial-intelligence-and-intellectual-property-copyright-and-patents-government-response-to-consultation
Intellectual Property Office (2024), Copyright and AI: Consultation (Intellectual Property Office, CP 1205). https://www.gov.uk/government/news/uk-consults-on-proposals-to-give-creative-industries-and-ai-developers-clarity-over-copyright-laws
JacksonF. (2025), ‘OpenAI and Google reject UK Government’s AI copyright proposal’, TechRepublic. https://www.techrepublic.com/article/news-openai-google-reject-uk-ai-copyright/
JasanoffS. & KimS.H. (2009), ‘Containing the atom: sociotechnical imaginaries and nuclear power in the United States and South Korea’, Minerva, 47(2): 119–146. https://doi.org/10.1007/s11024-009-9124-4
KahnG. (2025), How AI is Reshaping Copyright Law and What It Means for the News Industry (Reuters Institute for the Study of Journalism). https://reutersinstitute.politics.ox.ac.uk/news/how-ai-reshaping-copyright-law-and-what-it-means-news-industry
KL3M (2025), Webpage. https://kl3m.ai/
KleinbergJ. & RaghavanM. (2021), ‘Algorithmic monoculture and social welfare’, Proceedings of the National Academy of Sciences, 118(22): 1–7. https://doi.org/10.1073/pnas.2018340118
KrauseD. (2025), ‘The impact of low-cost AI: implications for big tech, market structures, and future growth’, in SSRN Scholarly Paper 5139411. https://doi.org/10.2139/ssrn.5139411
KretschmerM., MelettiB., BentlyL., CirfodelliG., EbenM., EricksonK., IraminaA., LiZ., McDonaghL., PerotE., PorangabaL. & ThomasA. (2025), ‘Copyright and AI: response by CREATe Centre to the UK Government’s consultation’, in CREATe Working Paper 2025/2. https://zenodo.org/records/14931964; https://doi.org/10.2139/ssrn.5166928
LandymoreF. (2025), Meta Says It’s Okay to Feed Copyrighted Books Into Its AI Model Because They Have No ‘Economic Value’ (Futurism). https://futurism.com/meta-copyrighted-books-no-value
LippmanD. (2025), An Explosion of AI Lobbying (Politico). https://www.politico.com/newsletters/politico-influence/2025/07/23/ai-lobbying-explosion-00472092
MansellR. (2012), Imagining the Internet: Communication, Innovation and Governance (Oxford University Press). https://doi.org/10.4324/9780367816742-18
MansellR. (2021), ‘European responses to (US) digital platform dominance’, in
JinD.Y. (ed.), The Routledge Handbook of Digital Media and Globalization (Routledge), 141–149.
MansellR. (2024), ‘Global communications governance research: colliding epistemologies and methodologies’, in
PadovaniC., WavreV., HintzA., GogginG. & IosifidisP. (eds), Global Communication Governance at the Crossroads (Palgrave Macmillan), 353–369. https://doi.org/10.1007/978-3-031-29616-1_20
MansellR. ‘Governing AI innovation under EU style capitalism’, in
BalčytienėA., SousaH., VanderbergheH. & Bajomi-LazarP. (eds), Power, Policies, and Governance of Digital Platforms: Expanding Implications for Democracy (NORDICOM) (in press).
MansellR. & PlantinJ.-C. (2022), ‘Imagining 5G networks: infrastructure and public accountability’, International Journal of Communication, 16: 4825–4843. https://ijoc.org/index.php/ijoc/article/view/18004
MansellR. & SteinmuellerW.E. (2022), ‘Denaturalizing digital platforms: is mass individualization here to stay? International Journal of Communication, 16: 461–481. https://ijoc.org/index.php/ijoc/article/view/13093
MansellR., DurachF., KettemannM., LenoirT., ProctorR., TripathiG. & TuckerE. (2025), Information Ecosystems and Troubled Democracy: A Global Synthesis of the State of Knowledge on News Media, AI and Data Governance (Paris, International Observatory on Information and Democracy). https://observatory.informationdemocracy.org/report/information-ecosystem-and-troubled-democracy/
MarshallJ. (2024), License or Sue: Tracking Publishers’ Relationships with Generative AI Companies. A Media Operator. https://www.amediaoperator.com/analysis/deal-or-no-deal-tracking-publishers-relationships-with-generative-ai-companies/
McGuiganL. (2023), Selling the American People: Advertising, Optimization, and the Origins of Adtech (MIT Press). https://doi.org/10.7551/mitpress/13562.001.0001
McKiernanJ. (2025), ‘Peers demand more protection from AI for creatives’, The Guardian. https://www.bbc.co.uk/news/articles/c39xj284e14o
McKinsey & Company (2024), Economic Potential of Generative AI (McKinsey & Company). https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier
McSherryC. (2023), Generative AI Policy Must Be Precise, Careful, and Practical: How to Cut Through the Hype and Spot Potential Risks in New Legislation (Electronic Frontier Foundation). https://www.eff.org/deeplinks/2023/07/generative-ai-policy-must-be-precise-careful-and-practical-how-cut-through-hype
MejiasU.A. & CouldryN. (2024), Data Grab: The New Colonialism of Big Tech (and How to Fight Back (WH Allen). https://doi.org/10.7208/chicago/9780226832319.001.0001
MiltnerK. & HighfieldT. (2024), The Possibilities of ‘Good’ Generative AI in the Cultural and Creative Industries (The British Academy). https://www.thebritishacademy.ac.uk/publications/the-possibilities-of-good-generative-ai-in-the-cultural-and-creative-industries/; https://doi.org/10.5871/digital-society/9780856726989.001
Mishcon de Reya (2025) Generative AI – Intellectual Property Cases and Policy Tracker. https://www.mishcon.com/generative-ai-intellectual-property-cases-and-policy-tracker
MokanderJ., BrockA., GriersonM., ZandermannK.L. & BradleyJ. (2025), Rebooting Copyright: How the UK Can Be a Global Leader in the Arts and AI (Tony Blair Institute). https://institute.global/insights/tech-and-digitalisation/rebooting-copyright-how-the-uk-can-be-a-global-leader-in-the-arts-and-ai
MontgomeryB. (2025), ‘AI companies start winning the copyright fight’, The Guardian. https://www.theguardian.com/technology/2025/jun/30/ai-techscape-copyright
Morgan Stanley (2023), The $6 Trillion Opportunity in AI (Morgan Stanley). https://www.morganstanley.com/ideas/generative-ai-growth-opportunity
MorrisR.G. (2022), ‘Selling out for a song: ‘artist abuse’ and saving creatives from servitude and economic disadvantage in the entertainment industry’, SMU Science and Technology Law Review, 25(2): 145–194. https://doi.org/10.25172/smustlr.25.2.3
MoyoL. (2024), Cultural Policy and Cultural Industries in Africa: From Culture as a Commodity to Culture as Praxis (Springer Nature). https://doi.org/10.1007/978-3-031-57742-0
News Media Association (2025), UK Creative Industries Launch ‘Make IT Fair’ Campaign Against AI Content Theft (Press Release). https://newsmediauk.org/blog/2025/02/25/uk-creative-industries-launch-make-it-fair-campaign/
NobleT. (2025), AI and Copyright: Expanding Copyright Hurts Everyone—Here’s What to Do Instead (Electronic Frontier Foundation). https://www.eff.org/deeplinks/2025/02/ai-and-copyright-expanding-copyright-hurts-everyone-heres-what-do-instead
NowotnyH. (2021), In AI We Trust: Power, Illusion and Control of Predictive Algorithms (Polity Press).
O’DonovanC., SarpG., WuX., StigloeJ., BertN., LiuS., ZamborskyT. & ZhaoT. (2025), Visions, Values, Voices: A Survey of Artificial Intelligence Researchers (University College London). https://zenodo.org/records/15118399
PiaseckiS. & HelbergerN. (2025), ‘A nightmare to control: legal and organizational challenges around the procurement of journalistic AI from external technology providers’, The Information Society: An International Journal, 44(3): 173–194. https://doi.org/10.1080/01972243.2025.2473398
PickardV. (2020), ‘Restructuring democratic infrastructures: a policy approach to the journalism crisis’, Digital Journalism, 8(6): 704–719. https://doi.org/10.1080/21670811.2020.1733433
QuijanoA. (2017), ‘Good living: between development and the de/coloniality of power’, in
RaussertW. (ed.), The Routledge Companion to Inter-American Studies (Routledge), 363–371.
RiederB. (2022), ‘Towards a political economy of technical systems: the case of Google’, Big Data & Society, 9(2): 1–5. https://doi.org/10.1177/20539517221135162
SavageM. (2024), ‘‘An existential threat’: anger over UK Government plans to allow AI firms to scrape content’, The Observer. https://www.theguardian.com/technology/2024/oct/26/an-existential-threat-anger-over-uk-government-plans-to-allow-ai-firms-to-scrape-content
ScottE. (2025), Copyright and Artificial Intelligence: Impact on Creative Industries (UK Parliament). https://lordslibrary.parliament.uk/copyright-and-artificial-intelligence-impact-on-creative-industries/
SchlesingerP. (2017), ‘The creative economy: invention of a global orthodoxy’, Innovation: The European Journal of Social Science Research, 30(1): 73–90. https://doi.org/10.1080/13511610.2016.1201651
SOA Policy Team (2024), The SOA Responds to the HarperCollins AI Licensing Deal (The Society of Authors). https://societyofauthors.org/2024/12/13/the-soa-responds-to-the-harpercollins-ai-licensing-deal/
Society of Audiovisual Artists (2025), Copyright and AI: 7 Recommendations to the UK (and the EU Alike) (Society of Audiovisual Artists). https://www.saa-authors.eu/en/news/938-copyright-and-ai-7-recommendations-to-the-uk-and-the-eu-alike
TaylorC. (2004), Modern Social Imaginaries (Durham, NC, Duke University Press). https://doi.org/10.1215/9780822385806
TehranianJ. (2025), The Secret Life of Copyright: Intellectual Property & Inequality in the Age of AI (Social Science Research Network). https://doi.org/10.2139/ssrn.5318606
ThomasA. & KretschmerM. (2025), The AI Licensing Economy (CREATe). https://www.create.ac.uk/blog/2025/02/24/the-ai-licensing-economy/#:∼:text=It’s%20unclear,predictable%20ways
TUC (Trades Union Congress) (2025), AI for Creative Workers: A TUC Manifesto (TUC). https://www.tuc.org.uk/sites/default/files/2025-02/AI_Manifesto_Creative_Workers_2025_AW.pdf
UK Government (1988), Copyright, Designs and Patents Act 1988. UK Public General Acts 1988 c48 as amended. https://www.legislation.gov.uk/ukpga/1988/48/contents
UK Government (2025a), Prime Minister Sets Out Blueprint to Turbocharge AI (GOV.UK Press Release). https://www.gov.uk/government/news/prime-minister-sets-out-blueprint-to-turbocharge-ai
UK Government (2025b), AI Opportunities Action Plan: Government Response (Dept for Science, Innovation & Technology CP 1242). https://www.gov.uk/government/publications/ai-opportunities-action-plan-government-response/ai-opportunities-action-plan-government-response
UK Government (2025c), Data (Use and Access) Act 2025. https://www.gov.uk/government/collections/data-use-and-access-act-2025
UK House of Commons (2024), Creator Remuneration: Fifth Report of Session 2023–24, HC156 (House of Commons Culture, Media and Sport Committee). https://committees.parliament.uk/publications/44143/documents/219382/default/
UK House of Commons (2025), British Film and High-end Television: First Report of Session 2024–25, HC 328 (House of Commons Culture, Media and Sport Committee). https://committees.parliament.uk/publications/47474/documents/246529/default/
UK House of Lords (2023), ‘At risk: our creative future—Communications and Digital Committee HL Paper 125’, in UK House of Lords, 2nd Report of Session 2022–23. https://publications.parliament.uk/pa/ld5803/ldselect/ldcomm/125/12502.htm
US District Court (2025a), Bartz et al. v. Anthropic Order on Fair Use, US District Court, Northern District of California, 23 June. https://assets.fenwick.com/documents/Bartz-v.-Anthropic-Fair-Use-Opinion.pdf
US District Court (2025b), Kadrey et al. v. Meta Platforms, Reply in support of Plaintiffs’ motion for leave to file third amended consolidated complaint, Case No. 3:23-cv-03417-VC. Northern District of California. 1 January. https://regmedia.co.uk/2025/01/10/pacer_kadrey_vs_meta_1.pdf
US District Court (2025c), Kadrey et al. v. Meta Platforms, Partial Summary Judgment. US District Court, Northern District of California. 25 June. https://www.courtlistener.com/docket/67569326/598/kadrey-v-meta-platforms-inc/
US Register of Copyrights (2025), Copyright and Artificial Intelligence: Part 2: Copyrightability (United States Register of Copyrights). https://www.copyright.gov/newsnet/2025/1060.html
van der VlistF., HelmondA. & FerrariF. (2024), ‘Big AI: cloud infrastructure dependence and the industrialisation of artificial intelligence’, Big Data & Society, 11(1): 1–16. https://doi.org/10.1177/20539517241232630
WalkerA. & VerseckaiteA. (2025), ‘AI, creativity, and copyright: finding the right balance’, techUK News. https://www.techuk.org/resource/ai-creativity-and-copyright-finding-the-right-balance.html
WattR. (2009), ‘An empirical analysis of the economics of copyright: how valid are the results of studies in developed countries for developing countries?’, in
WIPO (ed.), The Economics of Intellectual Property (World Intellectual Property Organisation), 65–108. https://www.wipo.int/edocs/pubdocs/en/economics/1012/wipo_pub_1012.pdf
WidderD.G. & HicksM. (2024), Watching the generative AI hype bubble deflate (arXiv:2408.08778). arXiv. https://doi.org/10.48550/arXiv.2408.08778
WilkinsonI. (2025), The EU’s New AI Code of Practice has its Critics but Will be Valuable for Global Governance (Chatham House). https://www.chathamhouse.org/2025/08/eus-new-ai-code-practice-has-its-critics-will-be-valuable-global-governance
WillemynsJ. (2025), Copyright & AI: The Case for a Pro-Growth Approach (Centre for British Progress). https://britishprogress.org/uk-day-one/copyright-ai-the-case-for-a-pro-growth-approach
WinogradT. (1990), Thinking Machines: Can There Be? Are We? In The Foundations of Artificial Intelligence (Cambridge University Press), 167–189. https://doi.org/10.1017/CBO9780511663116.017
ZuboffS. (2019), The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (New York, Public Affairs).
1
Social imaginaries of digital innovation are discussed in Cammaerts & Mansell (2020), Mansell (2012) and Mansell & Plantin (2022). In the field of Science and Technology Studies (STS), the term ‘sociotechnical imaginaries’ is used to refer to ‘collectively imagined forms of social life and social order reflected in the design and fulfilment of nation-specific scientific and/or technological projects’ (Jasanoff & Kim 2009: 120). Both concepts are concerned with normative perspectives and values. The former has been applied in institutional-level analysis; the latter often is more concerned with micro-level analysis and individual actors.
2
For example, in the UK a survey of AI researchers found that 51% of respondents agreed that innovation in the AI sector is inevitable (O’Donovan et al. 2025).
3
This forecast is influenced by global market projections, such as McKinsey’s estimate of US$2.6 to 4.4 trillion in gains from GenAI alone and Morgan Stanley’s projection that AI systems represent a US$6 trillion opportunity (McKinsey & Company 2024; Morgan Stanley 2023).
4
For example, authors in the UK earned an annual average of £7,000 in 2022, decreasing by 60% from 2006; the equivalent figure for visual artists in 2024 was £12,500, a 47% reduction from 2010 (Kretschmer et al. 2025).
5
In the US, the Pew Research Centre has reported that Google search users who see an AI summary are less inclined to click on links to other websites that users who did not see a summary in 2024 (Chapekis & Lieb 2025).
6
This assessment of press coverage is based on a review of policy debate coverage by three mainstream broadsheet newspapers in the UK (The Guardian, The Financial Times, The Times), press releases posted by three law firms and articles in four tech press outlets in 2024 to March 2025. A formal content analysis was not conducted, but 128 items were reviewed for their thematic content and perspectives on the issue of copyright reform and TDM.
7
This coalition included the British Phonographic Industry, the Independent Society of Musicians, the Motion Picture Association and the Society of Authors, as well as Mumsnet, The Guardian, The Financial Times, the Telegraph, Getty Images, the Daily Mail Group and Newsquest.
8
For example, ‘Baroness Kidron’s amendment, After Clause 132’, Data (Use and Access) Act 2025, Committee Stage, https://bills.parliament.uk/bills/3825/stages/19280/amendments/10016157.
9
Working groups have been established with participants from the creative industries and the AI industry (including the News Media Association, Alliance for IP, the Publishers Association and The Guardian, Sony Music Entertainment, Open AI, Amazon and Meta).
10
11
Kretschmer et al.’s (2025) analysis of 83 licences showed that 68% were with news or media companies, 14% with image banks, 7% with scholarly publishers, 7% with social media companies and 3% with music companies.
12
Seventeen companies were listed as certified as of August 2025, producing GenAI applications mainly for music, but also voice conversion and games.
Cite
DOI: 10.5871/jba/013.a38
Published on: 18 September 2025
Volume: 13
Issue: Issue 3
Article view count: 1094
Article download count: 84
Copyright statement
© The author(s) 2025. This is an open access article licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International LicenseCite this article
Mansell, R. (2025), ‘A capitalist contest: the AI industry v. the creative industries’, Journal of the British Academy, 13(3): a38 https://doi.org/10.5871/jba/013.a38Crossref
Dimensions
Altmetric
Article metrics
View count: 1094
Download count: 84
No Data Found
No Data Found
No Data Found
Received
October 2022
Accepted
May 2023
Published
July 2023
Voice conversion and cloning: psychological and ethical implications of intentionally synthesising familiar voice identities Carolyn McGettigan, Steven Bloch, Cennydd Bowles, Tanvi Dinkar, Nadine Lavan, Jonathan Chaim Reus, Victor Rosi
Voice identity conversion and cloning technologies use artificial intelligence to generate the auditory likeness of a specific human talker’s vocal identity. Given the deeply personal nature of voices, the widening availability of these technologies brings both opportunities and risks for human society. This article outlines key concepts and findings from psychological research on self-voice and other-voice perception that have a bearing on the potential impacts of synthetic voice likenesses on human listeners. Additional insights from speech and language therapy, human–computer interaction, ethics, and the law are incorporated to examine the broader implications of emergent and future voice cloning technologies.
11 September 2025
Democracy and discernment: the public purpose of the University in a world of ‘AI everywhere’ Katy Hayward
We are entering an era of ‘AI everywhere’, in which Artificial Intelligence is not merely being utilised but integrated into countless aspects of human interaction and occupation. Such is the nature of this integration that AI is not only changing our behaviour but our capabilities. It is also exposing new societal vulnerabilities—particularly in our use of language and the way we access, accumulate, and process information. Communication and information are all-important for the exercise of power, including political power. To hold such power to account, it is imperative that human intelligence remains distinctive and discerning. The inculcation of neoliberalism and endemic thoughtlessness has compromised the capacity of the contemporary University to meet this challenge. This article argues that the need for the University to foster the freedom of human intelligence is not only a moral obligation but an urgent democratic necessity.
23 March 2026
No Data Found
