UK Museums Search for a "Collective Voice" on Artificial Intelligence

Event: Let's Get Real: AI

Venue: The Audience Agency

Location: United Kingdom (national programme)

Dates: April 2025 - March 2026 (report published)

A new report from the Audience Agency has found that Britain's cultural sector needs to urgently develop a shared, coordinated approach to artificial intelligence, rather than allowing individual institutions to navigate the technology in isolation. The findings, published in March, come from an eight-month collaborative action research programme examining how AI is already reshaping museums, galleries and heritage organisations across the country.

Sixteen Organisations, One Shared Question

The report, titled Let's Get Real: AI, ran from April to December 2025, and brought together pairs of participants - typically a digital specialist and a senior manager - from 16 cultural organisations. Participating institutions included Bradford District Museums & Galleries, Bristol Museum, London Museum, The Box in Plymouth, and the National Gallery in London, offering a cross-section of the sector spanning major national institutions and smaller regional museums alike. The project was delivered in partnership with the Arts Marketing Association, Bloomberg Connects and the Institute for Digital Culture at the University of Leicester, led by its director, Ross Parry, alongside independent expert Jocelyn Burnham.

A Hands-On Methodology

The programme's methodology was deliberately hands-on rather than theoretical. Participants took part in two full-day in-person workshops, five online sessions exploring AI's ethical and practical dimensions, and five individual mentoring sessions in which each organisation designed, ran and evaluated its own small-scale AI experiment, tailored to its specific context. Rather than surveying opinion, the researchers analysed sixteen real, practical case studies, from which three recurring themes emerged: how staff felt about AI on a human level, how organisations approached experimentation and strategic guardrails, and how ready their technical and data infrastructure actually was.

Those three themes then formed the basis of the report's five core recommendations: keeping processes centred on people rather than allowing AI systems to dictate institutional priorities; establishing guardrails that remain flexible as the technology evolves; favouring small, low-risk experiments over large-scale rollouts; strengthening internal data practices before attempting to scale AI use; and actively sharing knowledge between organisations rather than working in isolation.

Opportunity Weighed Against Risk

Crucially, the report does not frame AI adoption as an unambiguous good. It explicitly weighs genuine opportunities against serious risks. As the Audience Agency put it in describing the sector's challenge, AI has real power to reshape how organisations engage audiences and demonstrate social impact - but, without ethical frameworks and skilled use, its adoption risks reinforcing bias, eroding public trust, and diverting effort away from mission-critical work. That tension between opportunity and risk runs through the entire programme, and distinguishes its findings from a simple advocacy case for AI adoption.

This caution carries particular weight for museums and galleries, given their position as trusted custodians of historical and cultural memory. Unlike a retail business adopting AI purely for efficiency, a museum's use of the technology touches directly on interpretation, representation and accuracy - issues that sit at the core of its public mission. A poorly considered AI tool used for collections research or visitor-facing content could distort historical narratives or introduce errors into information the public reasonably expects to be authoritative.

Finding a Collective Voice

As one contributor to the research put it, "the cultural sector needs to urgently find its collective voice and be part of the conversation". The phrasing captures a genuine anxiety within the sector: that decisions shaping how AI is regulated and integrated into cultural life are being made elsewhere, in technology companies and policy circles, without meaningful input from the institutions that will ultimately have to implement whatever framework emerges.

What Comes Next

The Audience Agency has signalled that this is not a one-off intervention. A follow-up programme, Let's Get Real: Social Impact + AI, launched on 25 June 2026 and is currently under way with a new cohort of cultural organisations, continuing until March 2027. Unlike the initial report, published in March, this second phase remains ongoing at the time of writing, with its findings and case studies not expected until a dedicated conference in spring 2027.

For now, the original report offers less a set of definitive answers than a structured starting point for a debate the sector clearly feels it cannot afford to delay. As museums, galleries and heritage organisations continue experimenting with AI in ways both visible and invisible to their audiences, the question at the centre of this research is not whether the technology will shape cultural institutions, but whether the sector will shape how that happens responsibly, rather than simply reacting to decisions made elsewhere.

Website

theaudienceagency.org