Stagely — Artificial Intelligence (AI) and Creativity Policy

Last updated: July 2026

At Stagely, we believe technology should widen access to theatre, for audiences, for the people who make the work, and for the people who run our business. AI is part of how we do that.

AI is a tool. It is not a substitute or replacement for human judgement or human creativity. Any tool can be used wisely or poorly, so it matters that we are open about how we AI tools witin our business and the Stagely app.

This policy sets out plainly where and how we use AI, where we don't, and the safeguards that keep human creativity at the core of our business.

1. Core principle: AI assists, humans decide

Every AI-assisted output that reaches a user, a theatre partner, or a funder passes through a human before it is published. We use AI to draft, summarise, sort, and suggest. We do not use it to decide.

The safeguards behind that:

• A named human signs off. Someone is accountable by name for every AI-assisted output we publish.

• Creative and editorial calls stay with people. What gets published, how a production is described, how a theatre or artist is represented, and what our product should be — these are human decisions, and AI does not make them.

• Nobody's creative work is displaced. We do not use AI in place of a writer, designer, photographer, or performer whose work we would otherwise commission or credit. Where a theatre-maker has produced the material, theirs is what we show.

• Mistakes are ours. Just like humans, AI can get things wrong. When AI makes a mistake, it is our mistake, not the tool's. We correct it, and we say so.

• The quick shortcut isn't always the right one. Sections 4 and 5 set out places where AI could plausibly save us time and money, and where we have chosen not to use it.

2. Accessibility and neurodiversity

Stagely was founded and is staffed by a team that includes neurodivergent people. We use AI openly as an accessibility tool, in the same way others use spell-checkers, screen readers, or dictation software.

• We use AI to help with reading dense material, structuring and redrafting written work, and organising information.

• This lets our team spend its energy on the parts of the work that need human insight — sector knowledge, relationships, product judgement, and creative direction.

• We regard this as levelling a playing field, not gaining an unfair advantage. The thinking, the evidence, and the decisions in our work are ours.

• Any document we produce with AI assistance is reviewed, edited, and signed off by a named human who stands behind its content.

3. How we use AI inside the Stagely app

Stagely brings together listings from thousands of professional and amateur productions across the UK. That volume requires automation. Here is how we use AI:

• Short production synopses. Where a listing has no usable description, or a description that is too long, inconsistent, or unclear, we may generate a short plain-English synopsis. These are drawn from source material published by the production or venue — not invented.

• Categorisation and tagging. We use AI to propose genre, format, and content categories so that productions can be compared and filtered fairly across very different source websites.

• Search and discovery. We use AI to interpret user searches and to match them to relevant productions, including regional and amateur work that would otherwise be hard to surface.

• Ingesting new productions. We use AI to help read and structure the highly variable data published across theatre websites and ticketing systems, so new productions can be added quickly and accurately.

• Show-level placeholder imagery. Stagely holds records at two levels: the show (the work itself, such as Macbeth) and the production (a specific staging by a specific company at a specific venue). Where a historic or frequently revived show has no single image that fairly represents the work, we may generate a neutral show-level image for it. We never generate imagery for a current production. Our rules on this are set out in section 4.

In every case:

• Output is reviewed by a member of the Stagely team before it is visible to users, or is flagged for review where confidence is low.

• Venues and producers can correct, replace, or remove any AI-assisted description of their own production, on request.

• Where a theatre supplies its own copy, that copy takes precedence. We do not overwrite a production's own words with machine-written ones.

4. Imagery: shows, productions, and where AI is used

Stagely distinguishes between a show — the work itself, such as Macbeth or The Importance of Being Earnest — and a production, meaning one company's staging of that work at a particular venue on particular dates. The distinction matters, because it determines what imagery we will and will not generate.

• We never use AI-generated imagery for a production. Production imagery represents the design, direction, casting, and photography of a specific creative team. It comes from the venue, producer, or rights holder, or we show none. A generated image has no business standing in for a company's work.

• We may use AI-generated imagery at show level, for historic works only. Many classical and long-out-of-copyright works — much of Shakespeare's canon, for example — have no single image that fairly represents them. In those cases we may generate a consistent set of neutral images tied to the work itself, so that a title appears in the app with something meaningful rather than a blank space.

Why generated images rather than the alternatives. The obvious substitutes are worse:

• Licensing a book jacket or publication cover attaches a publisher's design and interpretation to a theatrical work it has nothing to do with.

• Reusing photography from an earlier production credits one company's set, costume, lighting, and performances to a different company's staging. That is a misattribution of creative work, and we will not do it.

A neutral, show-level image avoids taking anyone's work and presenting it as somebody else's.

Our rules for show-level generated imagery:

• Follow a set format for identification. Generated show-level images follow a set format and can be easily identified.

• It never depicts real people. We do not generate images of identifiable performers, creatives, or venues.

• It does not imitate protected artwork. We do not generate images that reproduce or closely imitate the licensed key art, branding, or design associated with a title or its rights holders.

• It is consistent. Show-level images use a single house visual style, so they read as Stagely's own furniture rather than as any company's promotional artwork.

• Production imagery always displaces it. Where a production supplies its own image, that image is what a user sees for that production. The show-level image is a fallback, never an override.

• This is a temporary solution. Because we are a self-funded startup covering thousands of titles, it's a practical compromise, not our preference.

5. Where we do not use AI

• We do not generate reviews, opinions, or audience feedback. Anything presented as a human view on Stagely comes from a human.

• We do not use AI to replace the work of theatre-makers, designers, photographers, or writers, and we do not ask our suppliers or partners to do so.

• We do not use AI to decide representation or accessibility tags on its own. Tags such as LGBTQ+, ethnicity, D/deaf and disabled-led work, or relaxed and captioned performances carry real weight for the audiences who rely on them. These are confirmed by the venue or producer, or verified by a human, before they are shown. Getting these wrong causes harm; we would rather show nothing than show something unverified.

6. Accuracy, correction, and challenge

Automated systems get things wrong. Our commitments:

• Any user or venue can report an inaccurate listing, synopsis, or tag, and we will review it.

• Any venue or producer can ask to opt out of AI-assisted descriptions for their productions entirely.

• Where a rights holder or estate objects to a generated show-level image for a title, we will withdraw it.

• We keep a record of corrections and use them to improve our checking process, not just the individual entry.

7. Data, privacy, and public value

• We do not put personal user data into third-party AI tools for the purpose of training those tools.

• Our use of AI complies with UK GDPR and our published privacy policy.

• Audience insight we share with theatres is anonymised and aggregated. It exists to help venues understand and grow their audiences — a way of giving back value in return for the public funding that supported our development.

• We are transparent with partners about which parts of our platform are automated.

8. The tools we use, and how they are approved

The Stagely team uses AI tools regularly, but not indiscriminately. We maintain an internal list of approved tools, and team members use AI in their work only from that list.

How a tool gets onto the list. Before a tool is approved we consider:

• what it does with the data we put into it, and in particular whether that data is used to train the provider's models;

• whether it is covered by an enterprise or business agreement rather than consumer terms;

• where data is processed and stored;

• whether the provider's terms are compatible with UK GDPR and our obligations to users, theatre partners, and funders.

Keeping analysis inside our existing stack. Much of the Stagely platform is built on Google Firebase, so our default is to use AI tooling within that same environment — including Google Gemini — to interpret platform data and understand user trends. Keeping analysis inside the platform where the data already sits means fewer copies of that data in fewer places, under a single set of contractual terms, rather than moving it into unrelated third-party services.

How we handle data in these tools.

• Analysis of user behaviour is carried out on aggregated or pseudonymised data wherever the question can be answered that way.

• Personally identifiable user data is never put into an AI tool. Not in any tool, not in any tier, not for any purpose. This is an absolute rule, not a preference, and it applies exactly as set out in our privacy policy.

• This is enforced by system design, not just by instruction. Stagely staff work within a secure content management system that restricts access to registered user data. Staff cannot input personal data into an AI tool because our systems restrict access to that data in the first place.

• We do not put partner commercial terms or confidential third-party material into general-purpose consumer AI tools.

Keeping the list current. The approved list is reviewed at least annually and whenever we adopt a new tool or a provider materially changes its terms. Tools that no longer meet the criteria are removed.

9. Suppliers and partners

• We tell our partners and funders where AI is used in our work, as this policy does.

• We expect suppliers working on Stagely to follow the same principle: AI may assist, but a named human is accountable for what is delivered.

• We do not require partners to be AI-free, but we do reserve the right to reject or withdraw any material that misrepresents machine-generated work as human-made, or that infringes the rights of artists.

10. Review

This policy will be reviewed at least annually, and sooner if our tools, our product, or the regulatory landscape change materially. Our methods will evolve; the principle that a human is accountable for everything we publish will not.

Contact

Questions or feedback on this policy: daniel@stagely.app

Stagely Ltd, United Kingdom