Practical AI support for real cultural work
The AI Lab helps arts, culture and heritage organisations explore and use AI in ways that are useful, responsible and connected to real work.
It brings together practical use cases, prompts, workflows, assistants, experiments, templates and guardrails so organisations can understand where AI might help, test it safely and build confidence without getting lost in hype.
AI can help people research, organise information, develop ideas, analyse feedback, improve content and reduce repetitive work. But useful AI adoption is not simply about finding the latest tool.
It requires context, judgement, good processes and clear decisions about where AI should, and should not, be used.
The AI Lab is a place to explore those questions together.
Find useful opportunities
Start with the work, not the technology.
Explore practical ways AI might support tasks such as content development, campaigns, audience research, administration, fundraising, learning, collections, reporting, planning and internal knowledge.
The aim is to identify situations where AI can genuinely save time, improve quality or make something easier, rather than introducing AI simply because it is available.
Give AI better context
Generic prompts tend to produce generic answers.
Useful AI needs to understand the organisation, audience, objective, source material, tone, constraints and context around the task.
The AI Lab helps organisations create reusable briefs, profiles and structured context that make AI-supported work more relevant and consistent.
This also makes it easier for teams to understand what information they are giving an AI system and why.
Build repeatable workflows
A useful prompt is a good start. A reliable process is much more valuable.
AI Workflows show where AI fits within a wider task, what needs to happen before it is used, what information it needs, what a person should review afterwards and what happens next.
That turns isolated experiments into processes that can be understood, repeated and improved.
Test before scaling
Not every AI idea will be useful.
The Lab encourages organisations to begin with small, controlled experiments.
Try one task. Compare the AI-supported approach with the existing process. Look at the quality of the result, the time involved, the risks and the experience of the people doing the work.
If it helps, improve it.
If it does not, stop.
Small experiments provide better evidence for deciding where AI belongs within an organisation than adopting a tool across an entire team before understanding its value.
Keep human judgement
AI can support decisions and creative work. It should not remove responsibility for them.
The AI Lab builds review and approval into its approaches so that organisations consider accuracy, bias, privacy, confidentiality, copyright, accessibility, tone, provenance and the potential impact of an output.
Some tasks may be suitable for significant AI assistance. Others may only need AI at one stage. Some should remain entirely human.
The important thing is making that decision deliberately.
Learn as AI changes
AI is developing quickly, but cultural organisations do not need to follow every product announcement or experiment with every new model.
The AI Lab can help the community identify developments that are genuinely relevant, test them in cultural settings and share what is learned.
That makes it possible to keep pace with useful change without allowing the technology itself to set the agenda.
Practical AI support for real cultural work
The AI Lab provides different ways to explore, test and adopt AI depending on the task and the organisation’s level of confidence.
AI Use Cases
Practical examples showing where AI can support real work.
These might include:
- summarising audience feedback
- developing campaign ideas
- reviewing website content
- creating initial FAQ structures
- repurposing existing content
- analysing survey responses
- drafting internal briefing material
- preparing research summaries
- developing event communications
- supporting funding research
- organising organisational knowledge
- exploring collections information
- creating first drafts of plans or reports
Each use case should explain where AI adds value, what information it needs and what still requires human judgement.
Prompt Packs
Reusable collections of prompts for common tasks.
Prompt Packs provide a better starting point than an empty chat box and can include guidance on the context, source information and output format needed to get a useful result.
They can be adapted to suit an organisation rather than treated as magic instructions that always produce the right answer.
AI Briefs
Structured documents that provide AI systems with the background they need before completing a task.
An AI Brief might include information about the organisation, audiences, programme, tone of voice, objectives, constraints, terminology and relevant source material.
Creating that context once can improve many different AI-supported tasks.
AI Workflows
Step-by-step processes showing how AI can be used as part of a wider piece of work.
A workflow can explain:
- what needs to be prepared
- what information can be provided to AI
- which part of the task AI can support
- which prompts or tools to use
- what needs to be checked
- who should approve the result
- what happens afterwards
The focus is on creating dependable working practices rather than collections of disconnected prompts.
AI Assistants
Reusable AI setups designed around recurring tasks.
An organisation might develop an assistant for campaign planning, website review, meeting summaries, content repurposing, audience questions, policy research or another repeated activity.
Where similar needs exist across several organisations, Culture Hub can explore whether a shared assistant or framework can be created once and adapted by the wider community.
AI Experiments
Small tests designed to answer a specific question.
Can AI reduce the time it takes to analyse visitor feedback?
Can it help create better first drafts of exhibition FAQs?
Can it make a repetitive reporting process easier?
Can an existing piece of content be repurposed more efficiently without reducing quality?
Experiments help organisations gather evidence before deciding whether a particular use of AI should become part of normal practice.
AI Tool Reviews
Plain-English reviews of selected AI tools and features.
Reviews can consider what a tool does, who it might be useful for, how much it costs, what information it requires, how difficult it is to implement and what organisations should consider before using it.
The aim is not to create a race to adopt the newest technology.
It is to help organisations make informed choices.
AI Templates
Reusable structures that make AI work more consistent.
These might include:
- prompt templates
- AI briefs
- output specifications
- review checklists
- experiment plans
- workflow templates
- risk assessments
- approval records
- AI use registers
Templates reduce the need for every organisation to invent its own approach from scratch.
AI Guardrails
Practical guidance for using AI responsibly.
Guardrails can help organisations think about:
- accuracy and verification
- personal and confidential information
- copyright and intellectual property
- bias and representation
- accessibility
- transparency
- environmental impact
- tone and organisational values
- human review and approval
- appropriate record keeping
- situations where AI should not be used
The aim is not to prevent experimentation. It is to make experimentation thoughtful, proportionate and responsible.
AI Examples and Outputs
Real examples can make the difference between talking about AI and understanding how to use it.
The Lab can show before-and-after examples illustrating how vague instructions produce generic outputs, while stronger context, better source material, clearer workflows and human review create more useful results.
Examples can also show where AI has failed, introduced inaccuracies or simply not been worth using.
Those lessons are just as valuable.
From responsible guidance to practical adoption
Culture Hub does not need to invent its own principles for responsible AI in isolation.
Arts Council England, the Digital Culture Network and other sector organisations are already developing trusted guidance around AI policy, risk, readiness and responsible experimentation.
The role of the AI Lab is to help organisations turn that guidance into working practice.
A policy guide can become a Workbook.
A responsible AI checklist can become part of an implementation process.
A risk framework can become a practical review tool.
Guidance on running an AI pilot can become a shared experiment involving several organisations.
This allows Culture Hub to build on trusted sector expertise while concentrating on the gap between understanding what responsible AI looks like and actually putting it into practice.
Experiment together, learn together
Many cultural organisations are currently asking the same questions about AI.
Instead of every organisation experimenting privately and learning the same lessons separately, the AI Lab creates a place to test ideas together.
A small group of organisations might explore the same use case using different tools or approaches.
The community can compare:
- what worked
- what saved time
- what improved quality
- where problems appeared
- what risks needed managing
- what people thought of the process
- whether the approach was worth continuing
Those findings can then improve the workflow, prompts, guardrails and guidance available to everybody else.
AI should support culture, not flatten it
Cultural organisations hold knowledge, stories, collections, relationships and creative expertise that cannot simply be reduced to training data and automated outputs.
AI can be a useful tool within that work, but the distinctive value still comes from people.
Curiosity, lived experience, specialist knowledge, judgement, creativity, relationships and responsibility remain human.
The purpose of the AI Lab is therefore not to automate culture.
It is to help cultural organisations understand where AI can support their people, remove unnecessary work and create more space for the things that people do best.