A brief history of machines and storytelling

Automated storytelling long predates contemporary neural nets. Early experiments in generative text go back to rule-based programs and algorithmic poetry in the mid-20th century. In the 2010s, recurrent neural networks produced surreal short scripts, most publicly exemplified by the 2016 short film Sunspring, written with a neural net and staged as a proof of concept for machine-authored screenwriting. The last half-decade accelerated work: transformer-based language models (notably the GPT family and other large language models) dramatically improved coherence, style mimicry, and interactive responsiveness. Theatre-makers—always resourceful adopters of new tools—began experimenting with these models as creative collaborators rather than simple text generators. What began as curiosity quickly shifted into a practical set of dramaturgical techniques: ideation prompts, character-voice sketches, beat suggestions, and rapid dramaturgical outlining. The technology did not create theatre in a vacuum; it reframed long-standing practices of collaborative writing, tablework, and iterative workshop processes.

Tools, techniques, and the dramaturgical workflow

Contemporary dramaturg-AI workflows use a mix of general-purpose language models, tailored fine-tuning, and human editorial practice. Writers feed models scene prompts, character histories, or conflict summaries and receive multiple variants—dialogue turns, stage directions, or condensed scene synopses. Tools range from off-the-shelf chat interfaces to custom pipelines that include style conditioning, red-teaming for safety, and iterative human curation. The practical advantage is speed: what might take a writer hours to sketch can be expanded into multiple drafts in minutes, enabling more rapid exploration of tonal possibilities and structural permutations. Theatre practitioners have adapted techniques from software design—A/B testing of beats, prompt engineering as a rehearsal discipline, and version control for script iterations. In conservatories and community companies, AI is being used as a democratizing rehearsal partner: inexperienced writers can quickly prototype scenes; directors can audition multiple dramaturgical approaches before committing rehearsal time; and ensemble improvisors can prime machines with character backstory to generate provocative prompts. Importantly, the AI is rarely treated as author in its own right in most professional settings; rather, it functions as a creative accelerant and sounding board that surfaces options a human team can accept, refine, or reject.

Recent experiments and current news

By mid-2024, numerous small- and mid-scale experiments had moved from salons and lab reports into public-facing projects. Universities and labs continued to publish methodological papers demonstrating how controlled conditioning of models yields more reliable narrative arcs and character consistency. Fringe festivals and interdisciplinary showcases have featured pieces where AI-generated fragments were incorporated into live performance, provoking audience discussion about agency and authorship. Meanwhile, industry-level conversations have been dominated by labor and legal developments. The 2023 strikes by writers and performers in major English-language markets foregrounded the risk that generative AI could be used to train on copyrighted scripts without consent and to undercut writers’ bargaining power—concerns that continued to shape contracts and guidelines into 2024. Simultaneously, a growing number of theatre institutions began drafting internal policies about disclosure of AI use in development, and several grant programs started funding projects that position AI explicitly as a tool for co-creation rather than replacement.

Creative impact: what changes onstage and in rehearsal

Using AI as a dramaturg alters the texture of collaboration in several concrete ways. First, it shifts the tempo of experimentation: playwrights can iterate faster through tonal registers, enabling riskier formal play that would be expensive under traditional rehearsal constraints. Second, it introduces a deliberate indeterminacy—machines suggest possibilities that are not derived from lived experience, which can force creators to clarify motives and stakes more rigorously. Third, it amplifies voice mimicry: models fine-tuned on particular playwrights’ corpora can emulate cadence and idiom, a capability practitioners both admire for ideational breadth and debate for ethical implications. The emergent creative possibilities are tangible—hybrid works that interleave human-authored scenes with AI-suggested monologues, dramaturgical prompts that generate audience-interactive moments, and research-based plays that use AI to aggregate oral histories into composite characters. Reception so far has been mixed: critics and audiences often find AI-inflected pieces intellectually intriguing but sometimes emotionally uneven. The strongest responses tend to come from works that use AI deliberately as a provocation—integrating its limitations and biases into the dramaturgy—rather than masking it as invisible authorship.

Labor, authorship, and ethical stakes

Theatre is a profoundly collaborative, delivered-live medium, and introducing AI into authorship questions raises legal, ethical, and labor issues. Writers’ unions and performers’ groups have pushed for language that prevents unconsented replacement or undisclosed use of training materials. The core concerns are copyright (models trained on proprietary scripts), attribution (who is credited when AI contributes substantive material), and economic displacement (how AI-assisted drafting affects pay structures). Ethical frameworks emerging from arts organizations recommend transparency with funders and audiences, consent from collaborators when models are trained on their work, and contractual protections that preserve human decision-making authority. There is also an aesthetic ethics: relying on AI-generated emotional shortcuts can flatten lived specificity; hence, many dramaturgs argue for a pedagogy where AI serves to complicate—not simplify—moral and relational nuance. A parallel conversation centers on bias: models reflect the corpora they were trained on and thus can reproduce stereotypes unless actively countered in the editorial process.

Case studies and workshop examples

Several illustrative approaches demonstrate practical, responsible use. In one university program, student ensembles used models to generate alternative scene continuations during in-class table reads; students then performed multiple continuations to explore character choice. An independent company in a major city used an AI to suggest structural edits during a week-long writers’ residency, enabling the playwright to test three radically different act structures quickly; the final produced text was human-authored but owed its structural audacity to those rapid experiments. Festival programmers have safeguarded public trust by labeling works that used AI in development, which invited critical conversation rather than surprise. These case studies converge on a pattern: AI is most effective when treated as a catalytic tool within a transparent creative ecology that primes human judgment, archival sensitivity, and ethical oversight.

Looking forward: practices and policies that matter

If AI is to become an enduring part of theatrical craft, the sector will need a set of shared practices. First, disclosure norms: audiences and collaborators should know when AI has materially shaped a work, preserving trust and enabling informed criticism. Second, training and pedagogy: dramaturgy curricula must teach prompt design, model limitations, and bias mitigation. Third, contract practices: agreements should clarify ownership of AI-assisted drafts and protect people from replacement by synthetic text or voices. Fourth, curation and criticism: critics and programmers will need vocabularies to assess AI’s role in creativity, distinguishing bona fide innovation from gimmickry. Finally, continued public funding and philanthropic support for projects that experiment ethically will be crucial to ensure that AI tools benefit a diverse ecosystem rather than concentrating advantage in well-resourced companies. The generative question is not whether AI will alter playwriting—it already has—but how the field shapes those changes to protect craft, labor, and meaningful theatrical experience.

co-authorship as a creative choice

AI dramaturgy is not a single phenomenon but a constellation of practices that reconfigure rehearsal logic, authorship debates, and audience expectations. Historically, new tools—from lighting rigs to video projection—have been integrated into theatre with initial anxiety and eventual aesthetic richness. The current moment offers a similar inflection point: practitioners can treat AI either as a cheap imitation of human nuance or as a disciplined collaborator that surfaces options, complicates assumptions, and accelerates experimentation. The difference will lie in governance, pedagogy, and artistic intention. When AI is orchestrated transparently and critically, it can expand the dramaturgical palette without erasing the human act of taking responsibility for what happens onstage.