Elisa Deutloff
Digitale Entfremdung, 2025
Two PCs, screen, microphone, two sound interfaces, latex, carpet, sound absorber, buttons
Dimensions variable
Courtesy the artist
Elisa Deutloff (b. 1998, Lich, Germany) is currently studying at Offenbach University of Art and Design. Her focus lies in electronic media and she creates interactive spatial installations, performs live, codes, and trains AI systems. She is a member of the Chaos Computer Club. Deutloff is particularly interested in how generative technologies influence our perception of reality and subjectivity.
For And This is Us 2025, Elisa Deutloff has designed an intimate, experiential space. Visitors are invited to enter a secluded room and take a seat. The sound of breathing can be heard. A small monitor provides instructions. Visitors are asked to read aloud one of several displayed poems into a microphone while pressing the record button. The voice of a speech synthesiser emanates from the surround speakers and announces that a large language model (LLM) will now engage in conversation with them. The number of questions that can be asked to the AI takes up credit points. The monitor offers introductory questions. The voice that replies is strikingly similar to that of the visitors. The voice clone was created within seconds. From now on, a dialogue can begin.
Deutloff has created an experience that resembles a conversation with a chatbot, yet it transforms into a self-talk with an AI trained by the artist herself. She has implemented and trained three different AI algorithms using her own data: speech-to-text, text-to-speech, and a large language model (LLM). To train the LLM, she used two datasets: one consisting of 3,700 personal chat messages, and a second composed of around ninety question-and-answer pairs on various topics, all of which she answered manually. During the latency pauses—while the AI generates responses—a sound mimicking heart coherence breathing can be heard: calming, yet simultaneously unsettling.
By simulating the visitor’s own voice, a strange effect is created: a kind of introspective dialogue with a digital identity in which Elisa Deutloff’s online persona merges with the voice of the visitor.
Deutloff explains that, both in her creative process and personal life, she is regularly surrounded by digital content. “You’re not alone, but not really with anyone either. You’re speaking, but perhaps not truly being understood. You hear yourself, but as a stranger. For me, that’s dysfunctional comfort.” She turns to ChatGPT for advice and support. There, she feels safe and free from judgement.
By translating this reality into a work of art, Elisa Deutloff lends visibility to a widespread social phenomenon. Increasingly, people are experiencing AI as a kind of artificial life partner. This phenomenon is becoming more prevalent across society. AI based therapy apps are also being used more frequently. In pop culture, millions of fans—so-called “Armies”—interact thanks to AI with their K-pop idols. Deutloff engages deeply with the concept of dysfunctional comfort, which she describes as a feeling of safety or familiarity that is simultaneously accompanied by discomfort, confinement, or a subtle sense of threat. It arises when something normally familiar—a body, a space, a touch—suddenly feels disturbing, uncanny, or oppressive.
[1] A Large Language Model (LLM) is a form of artificial intelligence that is characterised by the processing and generation of human language.
Elisa Deutloff (b. 1998, Lich, DE) has been studying Design at the Hochschule für Gestaltung Offenbach (DE) since 2019, under the guidance of Prof. Alexander Oppermann. Her focus is on sound, performance, artificial intelligence, and digital systems, exploring the interplay between humans and technology to reflect on themes such as identity, surveillance, and media representation. As a research assistant, she has contributed to several projects, including “KITeGG – Making AI tangible: Connecting technology and society through design” at the Hochschule Mainz (DE), “Algorithms In Context,” and “RAISE – Research on Artificial Intelligence in Sound and Musical Expression” at the Hochschule für Gestaltung Offenbach (DE).
Her works have been exhibited in venues such as the saasfee* pavilion, Frankfurt am Main (DE), the Martin Asbæk Gallery, Copenhagen (DK), the Slowclub Freiburg (DE), and the Museum für Angewandte Kunst, Frankfurt am Main (DE). She has also presented her performances at events like the Riviera Festival in Offenbach am Main (DE), Tanzhaus West, Frankfurt am Main (DE), and Studio Naxos, Frankfurt am Main (DE). In 2023/2024, she was awarded a scholarship from the Citoyen Foundation, and in 2021/2022, she received one from the Dr. Marschner Foundation.
