Franz Hohler Tests AI Impersonation: Audience Fails to Spot Human vs. Machine

Swiss writer Franz Hohler conducted an experiment at the University of Lucerne to test whether audiences could distinguish his original writing from an AI-generated imitation. He presented two short stories—one written by himself, the other by ChatGPT in his style—to a live audience without revealing which was which. The AI text, which featured a protagonist named Herr Bärtschi and included a Swiss German expression ('im Tram'), was attributed to Hohler by about three-quarters of the audience, while his own story was mistaken for a machine's output. Hohler, known for his literary and cabaret work, expressed shock and compared the experience to Garry Kasparov's first chess defeat by a computer. He criticized the AI's text as a 'pseudo-poetic imitation' and a parody, yet acknowledged the difficulty of convincing the audience of its flaws. The experiment was part of a series on artificial intelligence at the university, where Hohler and young author Nelio Biedermann discussed their views on AI. Hohler's account of this experience will appear in his new book 'Der Schrei des Elektrikers', a collection of about five dozen speeches and texts honoring literary companions from Gerhard Meier to Peter von Matt, to be published in November by Luchterhand-Verlag. The event highlights the growing challenge AI poses to human creativity and the blurred lines between original and generated art.