Good decisions rely on good data, and without trustworthy data, insights (human or AI-supported) risk being shallow, biased or less impactful than they could be.
The Data Trust: Foundations for an AI-ready future
Together with thousands of contributing organisations and partners, The Audience Agency (TAA) has spent over a decade stewarding the Data Trust for the cultural sector – a shared, anonymised evidence base built from millions of records.
This collective effort provides the foundations of a perfect recipe, combining Data4Good and Tech4Good to ensure that AI4Good is grounded in evidence that is representative, ethical and sector-owned. By contributing and collaborating through the Data Trust, organisations and peers can share insights, protect values and co-create the tools the sector needs.
What follows sets out the ingredients of a recipe: how Data4Good underpins the Trust; how Tech4Good makes it usable, efficient and scalable; how People4Good bring meaning; and how AI4Good can then be introduced responsibly, ethically and with confidence.
Data4Good and Tech4Good: The Data Trust a decade of collective evidence
Data4Good is especially important for the cultural sector and the adoption of AI. Large Language Models (LLMs) such as those from OpenAI, are trained on vast swathes of internet data. Still, they do not capture the specific realities of our audiences, programmes and policy contexts. The result is that, while they may sound convincing, their answers can be inaccurate, misleading or irrelevant to the needs of cultural organisations.
The Data Trust is a shared, anonymised evidence base built from millions of records contributed by thousands of organisations. It acts as both a safeguard and an enabler, protecting the cultural sector from the risks of relying on ungoverned or commercially driven data, while creating the conditions for everyone to benefit from future innovations.
Thanks to all the organisations and partners involved, it has become the world’s largest anonymised cultural dataset, spanning over five million performances and 323 million tickets from venues, museums, festivals and producers across the UK. Each record is strictly anonymised, checked for quality and standardised to protect sensitive data.
As technologies like AI evolve, eco-systems like the Data Trust can support its introduction on the sector’s terms, grounded in evidence that is representative, ethical, and collectively owned. In this way, the sector can explore change with confidence, knowing the benefits are shared and the values of trust, fairness and collaboration remain at the centre.
The Data Trust is hosted on a robust, purpose-built data architecture, integrated into partners’ ecosystems, and refreshed over the last two years, which turns contributions into insights for those who share their data. Each feed is anonymised, checked and standardised. A collective effort where Tech4Good can scale the benefits of the Data Trust across the UK.
People4Good: Wisdom and collaboration
Of course, the most important ingredient in a Data Trust is people. At TAA, we collaborate with organisations to interpret evidence, test hypotheses against national/local knowledge, and translate findings into actionable insights, leveraging the expertise of practitioners who know their audiences best. Together, we turn data into impact.
Equally important is listening directly to communities. Alongside quantitative data, complementary qualitative insights are captured through conversations, surveys and co-created research. By involving communities in shaping evidence and its interpretation, insights are not only statistically robust but also grounded in lived experience. This helps organisations apply findings in ways that are meaningful, relevant and trusted by the people they serve.
From our experience of AI, we know success comes from carefully curating its use with people and not relying on generic AI tools. Helping organisations is so important – not only to make the most of data, but also to build confidence in their wider digital adoption, including AI. Our flagship programme, Let’s Get Real: AI, creates space for professionals to explore AI tools, governance, ethics and strategy in a peer-supported environment.
AI4Good: Ethical AI and our approach to it
The Data Trust continues to be developed so it is future ready for the adoption of AI with clear, ethical and transparent principles. These are built on four cornerstones:
- Ethics – Why are we doing this? AI must serve our mission and stakeholders and be fair to people and creators.
- Responsible use – Are we doing it right? We assess bias, privacy, IP and sustainability before use.
- Safety – What could go wrong? We maintain human oversight and higher risk uses undergo an AI Impact Assessment.
- Governance – Who’s responsible? Governance ensures AI is not adopted ad hoc, but through shared, transparent decisions that reflect sector values.
We apply these principles through explainable models, fairness checks and by considering the environmental impact. This ensures AI supports, rather than replaces, human judgment and creativity.
AI4Good: Data Trust innovation to support the sector
Through the InnovateUK BridgeAI programme, the first AI innovation in the Data Trust focused on Event and Artform Categorisation Tagging – a process that was previously manual and often inconsistent. With guidance from a scientific advisor at the Alan Turing Institute, we worked through a complex machine-learning challenge, comparing models, weighing costs and ethics, and ensuring the approach was transparent and fair.
The result was a system that classifies each performance record into a consistent two-tier taxonomy, always backed by human review. This is being rolled out across all five million performances in the Data Trust, unlocking consistent, national-scale benchmarks. You can attend the free Audience Trends Across Event & Artform Categories: the past five years for a first look.
The Data Trust innovation continues with the piloting of curated custom Audience Spectrum AI Agents to make guidance more interactive. AI Agents are like tailored digital assistants that utilise our trusted tools and knowledge, to quickly and safely answer sector-specific questions. The new AS Genie Beta helps organisations explore audience segments, motivations and strategies through a conversational interface. We are taking our time with these curated tools and working with organisations to test their application – please do reach out if you are interested.
Finally, the Data Trust, will benefit from what may be the world’s first Cultural Model Context Protocol (MCP) Beta server, a universal connector between AI tools. This step opens up future possibilities for organisations, funders and partners to collaborate, share data safely, and co-create the next generation of benefits for the sector.
Better data – better tools – better decisions
The Data Trust has already demonstrated the value of collective effort, and this same spirit of collaboration underpins the work of the National Cultural Data Observatory (NCDO) scoping project*.
The NCDO is laying the foundations for a stronger, joined-up cultural data ecosystem – one shaping long-term strategy for the sector while supporting day-to-day practice. In doing so, it helps overcome fragmented data and builds a coherent, defensible, and people-centred evidence base for the future.
Only with trusted, collective evidence can we combine the ingredients of Data4Good, Tech4Good, People4Good and AI4Good into a recipe that enhances strategy while keeping trust, fairness and creativity at its core.
The Data Trust provides a foundation, and by working together we can ensure that the benefits of AI are introduced on the sector’s terms – evidence-based, ethical and deeply human.
Contact us to join the conversation and shape what comes next.
*Led by the Centre for Cultural Value at the University of Leeds and funded by the ESRC, the project brings together partners including The Audience Agency, MyCake, Culture Commons, the University of Sheffield, Leeds Institute for Data Analytics, and De Montfort University, with the support of national bodies such as DCMS, Arts Council England, the Local Government Association, Historic England, and Bradford 2025.
This article was originally published on Arts Professional on 16 September 2025
