Molière Ex Machina: when AI lends its pen to the 17th Century

Supported by Sorbonne University, the Théâtre Molière Sorbonne and the Obvious trio, in partnership with Mistral AI, this interdisciplinary project employs an original creative process that combines the humanities, technology and traditional craftsmanship.

The full production will be performed at the Royal Opera at the Palace of Versailles on 5 and 6 May 2026.

What if Molière hadn’t died in 1673? This daring, almost sacrilegious idea gave rise to “Molière Ex Machina,” the result of an unexpected alliance between the classical humanities and artificial intelligence. The story began in 2023 during a round-table discussion on artistic creation and AI at Sorbonne University, where Pierre-Marie Chauvin, Vice-President for Arts, Sciences, Culture and Society, met the artists from Obvious. “This pioneering collective in generative art already had many connections with the university’s AI ecosystem, notably with ISIR and SCAI, but had not yet connected with the humanities and arts,” he explains. He introduced them to Georges Forestier, founder of the Théâtre Molière Sorbonne, a theater that uses reliable sources to accurately reconstruct 17th-century productions, including text, music, sets, costumes, and staging. This meeting sparked a fruitful curiosity between two worlds that initially seemed opposed.

The idea began to take shape: imagine the play Molière might have written had he lived another year. The theme of astrology naturally emerged. Present through minor characters in Molière’s work, as well as in his two personal libraries, astrology was nevertheless never directly addressed by the playwright. The play’s title, L’Astrologue ou les Faux Présages (The Astrologer or False Omens), suggested by the AI itself, invites us to reflect on our own contemporary fascination with predictive systems. “Astrology allows us to discuss manipulation, false beliefs and disinformation, which are particularly topical subjects,” emphasises Pierre-Marie Chauvin. Far from being a mere pastiche, the project becomes a critical mirror of our times, creating a dialogue between the false omens of the astrologers of the Great Century and the often opaque promises of contemporary digital technologies. Through an experience of speculative fiction, it becomes possible to examine our present through a past recreated by the machine.

The Théâtre Molière Sorbonne and Obvious joined forces with many experts, including classical literature scholars, linguists, theater historians, digital artists, actors and AI researchers. They are all part of the same research-creation process. “The idea was to see what generative AI could do by bringing the best people together and using a rigorous scientific approach,” says Pierre-Marie Chauvin. At the beginning of the adventure, several discussions took place with members of the Théâtre Molière Sorbonne, including Georges Forestier, a renowned Molière specialist. His death in 2024 brought the project to a standstill. The team decided, however, to continue, and to sustain the intellectual and artistic momentum he had helped to spring in motion.

In this project, artificial intelligence was never considered an autonomous author. “Our goal is to create with AI by demystifying it,” says Hugo Caselles-Dupré, a member of the Obvious trio. “When AI is well-guided and supervised, it becomes a creative partner.” Several approaches were tested: AI training using 17th century texts, guided generation based on excerpts from Molière’s plays, and direct exchanges with the model. The dialogue-based method proved the most effective. 

During long working sessions, the Molière Ex Machina team exchanged with the AI model as if it were a real writing partner. Situations were suggested, generated lines analysed, and new variations requested. Each scene went through numerous rewrites. The creators explained to the AI model why some proposals did not work, suggested dramatic devices typical of Molière’s comedy, and gradually refined the text. AI made suggestions; the artists selected, guided, and corrected them. “The process was long and demanding,” says Mickaël Bouffard, director and head of the Théâtre Molière Sorbonne. “Some scenes had to be revised multiple times.” 

But the project wasn’t limited to writing. It also drew upon the expertise associated with 17th century theater to design the costumes, build the sets, and write the music. Image generation models were trained using drawings by Henri Gissey, the costume designer for the court of Louis XIV, to produce visual proposals in the period’s style. The visuals were collectively discussed and adapted to ensure consistency with the theatrical practices of the era, and then served as the basis for the costume designer and the team of tailors. The same approach was adopted for the sets, drawing inspiration from Pierre-Paul Sevin’s drawings to produce sketches faithful to 17th century scenogaphic aesthetics. The music followed a similar path, with the AI model having been trained on works by Marc-Antoine Charpentier to generate scores in the Baroque style. 

More than a technological experiment, Molière Ex Machina offers a new way of bringing France’s literary heritage to life and passing it on. By bringing together university research, contemporary creation and classical theater, the project invites us to imagine what Molière’s work can still inspire today.

The next steps

With the play’s script now finalised, the first performances are scheduled to take place at the Royal Opera of the Palace of Versailles on 5 and 6 May 2026. The production is then set to tour Paris and other cities in France, and possibly abroad, with a number of venues already under consideration.

Educational activities will accompany the performances to raise awareness among different audiences of the project’s artistic, scientific and technological dimensions. 

The project as a case study

Contrary to the alarmist rhetoric concerning AI, “Molière Ex Machina” offers an example of a concrete and profoundly human experience in assisted creation, void of naivety or magical belief in technology. “We are demonstrating in concrete terms something that can be achieved in a novel way with AI. Not a play written by AI, but a play co-written with it.” The project thus aims to demystify AI, and to show how it can be used prudently and meaningfully, provided it is supervised, guided and informed by human knowledge. Pierre-Marie Chauvin emphasises, “In this play, which could not have come into being without AI, not a single word was written directly by a human. Yet nothing in the play is devoid of human supervision.” 

AI does not replace the author: it becomes one of the tools in the service of research-based creation. In this sense, “Molière Ex Machina” fully embodies the university’s core missions: to study the world, to share knowledge, and to foster collective reflection on today’s issues.