Kenneth Glad
At the Royal Opera at the Versailles, audiences watched a new Molière-style comedy whose most unusual collaborator was an AI system. The project has become a test of how far technology can enter literary heritage without replacing human authorship.
More than 350 years after Molière’s death, researchers in France have staged a new play developed with the assistance of artificial intelligence in the style of the 17th-century dramatist.
According to The Guardian, L’Astrologue ou les Faux Présages debuted at the Royal Opera at the Château de Versailles after a long collaboration between Sorbonne University, Théâtre Molière Sorbonne and the artist-researcher collective Obvious.
The production was not presented as a lost Molière text or a recovered manuscript. It was framed instead as an experiment:
What might happen if scholars, artists and an AI system tried to build a new comedy from the patterns, concerns and theatrical language associated with one of France’s most revered writers?
That question carries particular weight in France. Molière is not simply a historic playwright. His name is often used as shorthand for the French language itself, and his comedies remain central to the country’s literary and theatrical tradition.
Familiar comic machinery
The play’s plot draws on elements that would not feel out of place in 17th-century French comedy: A family trapped in a marriage scheme, a gullible man of status, a young woman whose future is being negotiated by others, and a supposed expert who uses false authority to manipulate those around him.
According to The Guardian, the story involves a wealthy Parisian father who is pushed by a charlatan astrologer toward a disastrous match for his daughter. 20 Minutes identified the father as Géronte and the astrologer as Pseudoramus, with the daughter Lucile caught in the middle.
The essential conflict can be understood broadly: Superstition and social ambition collide with love, money and deception. The false astrologer’s power depends not on magic, but on the willingness of others to believe him.
That theme gives the project a contemporary edge. A story about false predictions, manipulated belief and misplaced trust naturally invites comparison with modern anxieties about misinformation and automated authority.
Shaped with AI
According to the outlets, the French AI tool Le Chat was used throughout a two-and-a-half-year process involving theatre scholars, linguists, historians and researchers.
The project was not a matter of pressing a button and staging the output. The team reportedly carried out thousands of exchanges with the system, feeding it material, challenging what it produced and asking for revisions when scenes failed to meet the standards of the project.
Director Mickaël Bouffard said, writes the British newspaper, that the first AI version was only eight pages and “not very interesting.” He said scenes had to be revised repeatedly before the text could work on stage.
Bouffard also said: “AI has a superpower: The ability to store everything Molière wrote and everything Molière read. We human beings can’t do that.”
The remark points to the central bargain of the experiment. The AI could hold an immense archive in memory and produce variations at speed. But the judgment regarding tone, structure, theatrical usefulness and historical credibility remained with people.
Old stagecraft
20 Minutes reported that the production also drew on reconstructed 17th-century performance traditions, including period pronunciation and declamation. Screens in the hall displayed the dialogue, helping the audience follow the unfamiliar delivery.
That detail matters because the experiment was not only literary. It was also theatrical and historical. The play was placed inside a broader attempt to revive older stage practices rather than simply dress a modern script in antique language.
The involvement of Théâtre Molière Sorbonne gave the work an academic frame. The institution has focused on historically informed performance, making this project less like a commercial AI stunt and more like a controlled test of imitation, scholarship and theatrical craft.
The costumes, music, acting style and spoken delivery all helped determine whether the new text could survive in the world it was trying to evoke. A Molière-style script without that stage language would have been a different experiment entirely.
Critics respond
Reaction was mixed, though several early responses suggested that the illusion worked better than many skeptics might have expected.
Writing for 20 Minutes, journalist Christophe Séfrin called the AI imitation “striking, almost disconcerting” and said the dialogue was “entirely believable.”
His account emphasized how quickly the ear adjusted to the recreated language and how strongly the performance evoked the feel of a genuine period comedy.
Télérama described the project as a bold theatrical experiment, while also noting that parts could feel closer to pastiche than revelation.
That tension sits at the heart of the whole undertaking: Is the result a new artistic work, an academic reconstruction, a sophisticated imitation or some hybrid of all three?
Audience reaction, as reported by The Guardian, also reflected that divide. Some viewers found the result persuasive as theatre. Others were less convinced that artificial intelligence had added anything a skilled human playwright could not have achieved.
Artists and algorithms
The debate around L’Astrologue ou les Faux Présages extends beyond one performance in Versailles. It lands in the middle of a wider argument about whether generative AI should be treated as a creative tool, a threat to cultural labor or both.
A report submitted to the French National Assembly warned that AI may open new creative routes while also raising risks for jobs, rights and cultural diversity. The report called for a balance between technological development and the protection of artists and rights holders.
That concern echoes beyond theatre. The Guardian has separately reported that translators across Europe are already seeing machine-assisted work alter their profession, often through lower-paid editing of AI-generated drafts.
The link is not accidental. Theatre, translation, music, design and publishing all depend on forms of human skill that can look deceptively easy once a machine produces a fluent surface.
The danger, critics argue, is not only that AI may fail, but that institutions may decide that “good enough” output is sufficient.
A supervised experiment
The Sorbonne project’s defenders appear to be drawing a careful line. They are not saying that Le Chat has become a playwright in the human sense. They are arguing that AI, placed under expert supervision, can become part of a creative process.
Sorbonne academic Pierre-Marie Chauvin told The Guardian that the result was “not a play written by AI, but a play co-written with it.”
That distinction may determine how the work is remembered. If viewed as a replacement for Molière, the experiment is easy to reject. If viewed as a scholarly performance laboratory, it becomes more interesting: A way to test what style is, how much of it can be modeled, and where human taste still has to intervene.
For now, L’Astrologue ou les Faux Présages stands as a provocative theatrical case study. It does not settle the argument over AI and art, but it gives that argument a stage, a cast and an audience.
Sources: The Guardian, 20 Minutes, Télérama, French National Assembly report.
