This blog is the first of a series of pieces from JB on the topic of AI and its use in the creative industries. We will be exploring the environmental and ethical concerns arising from the use of this technology, how the sector is responding to these challenges, and what the future of AI in the arts might look like. This piece focuses on the ethical considerations of AI use.
In February last year, over 1,000 UK musicians released a collaborative album titled ‘Is This What We Want?’. Listeners rushing to press play were not met with creative harmonies or new, experimental alliances, but 12 tracks of near-total silence. The project was a protest against proposed changes to the UK copyright law that would have allowed companies developing artificial intelligence (AI) models to use copyrighted work of artists without permission or payment. To the musicians involved, the silence represented the devastating impact this law could have on artists’ livelihoods.
A month before the album was released, the UK Government published the AI Opportunities Action Plan, framing AI adoption as a key driver of economic growth and employment in the UK. This was shortly followed by a report from the National Engineering Policy Centre, which highlighted the significant environmental risk posed by AI. It called for mandatory reporting on water and energy use and carbon emissions by data centres in response to mounting evidence that the infrastructure underpinning AI carries serious ecological and social costs.
These opposing narratives point to a complex reality. AI – both in its traditional and generative forms, the latter of which can create new content from existing data – is embedded across different aspects of our lives. While there might be conversations about the AI ‘bubble’ bursting, it’s clear that it won’t be going away anytime soon. The appeal is the promise of more and better of seemingly everything: in our personal and professional lives, we’ll know more, we’ll do more, and have more time as a result of the efficiencies AI affords. But at what cost and to what end?
For the arts and culture communities, which are often centred around care, justice, and human-centred practices, the dangers of AI appear to challenge fundamental values while also potentially undermining the long-term viability of creative livelihoods through copyright infringements and unauthorised scraping. The creep of AI into every part of our lives should not be taken as an inevitable evil, but instead force us to question what it would look like to use AI ethically on an already-strained planet.
At JB, we’ve been thinking about how AI shows up not only in our day-to-day work, but also in the work of the artists and organisations we collaborate with. At the end of last year, some of the JB team ran a session for artists on AI to assess outlooks and demystify some of the uncertainties around its impact and scope. Feelings about AI were mixed. There was a clear consensus that AI could be used as a tool to enhance human capability, expand opportunity and make essential work more efficient and accessible, but also that significant risks were involved, including artistic job losses and a drift away from critical thinking in favour of shortcuts to meet unrealistic budgets and deadlines.
The ethics of AI
In recent months, much of the discussion around the ethical impacts of AI has, rightly, been focussed on its use in military operations. Reports of AI-enabled weapon systems being used – for example as part of the ongoing genocide in Gaza – have been covered across the news, prompting important examination of the intentions and ethics of large tech companies and serious questions around moral responsibility.
Beyond its use, there are also ethical implications surrounding how this information is generated in the first place. When we use AI tools, the results are often so fast, coherent and frictionless that there’s little time to think about where the information has come from or how it has been generated. But behind each answer, every suggestion or generated image, there’s information which was at some point created by someone, somewhere. It is worth being precise about how this works, because the picture is more complex than it first appears.
The training data feeding large AI learning models draws heavily on open web sources such as Wikipedia, GitHub, and other publicly available creative and academic work, raising urgent questions about intellectual property and consent. Separately, Big Tech companies employ workers in the Global South in content moderation and data labelling roles: work that is often psychologically harmful and poorly paid relative to the value it generates. Research published by the LSE documents the precarity and trauma embedded in this labour.
Increasing the transparency of how AI models are built will be key to building a more ethical future for the technology centred on respect and consideration for the work of others – those who have created the original research or creative works, and those feeding this information into AI models.
There is a further dimension that receives less attention: the language and cultural bias baked into AI systems from the start. Research has shown that training data is overwhelmingly sourced from English-language, Global North contexts. This means AI systems reflect and reinforce particular worldviews, marginalising other languages, knowledge and cultural frameworks – a concern for any organisation working across international communities and non-Western artistic traditions.
The rise of generative AI tools has also forced us to think about how we want to create and preserve originality. The arts are a vital way to connect with each other and with our shared humanity. When AI acts as an intermediary, how is that threatened? For a community often driven by the need for people-centred solutions, does using AI to generate answers represent a shortcut that comes at a cost? Or does it open up new ways of thinking and connecting? We need to create forums to explore these questions in an open and honest way.
Tools like Glaze, software designed to protect artists’ intellectual property by cloaking their work from AI learning models, point to creative and technical forms of resistance that are already emerging. Projects like Lelapa AI’s InkubaLM – a small language model built with and for low-resource African languages – show what AI development looks like when it is grounded in community, context and cultural specificity.
What we’re thinking about, and what we’d like to explore with you
At JB, we’ll be continuing to explore how AI shows up in our work and how we ensure that, where it does, it aligns with the values of care and climate justice that sit at the heart of our organisation and many of our partners. Rather than viewing it as an unstoppable technology over which we have no control, we’ll draw inspiration from projects that imagine otherwise and develop approaches to AI that are both ethical and ecologically sound.
Spread the word