Individuals are increasingly turning to generative artificial intelligence, particularly chatbots powered by large language models (LLMs), to support their decision-making. These systems answer questions in real-time, provide information and even offer specific moral guidance. When it comes to ethical advice, LLM-based chatbots are often seen as a double-edged sword: they hold promise for supporting users in navigating complex decisions, but also raise concerns about their impact on human moral reasoning and the ability to think critically.

To better understand and proactively shape the role of AI chatbots in ethical decision-making, it is essential to involve stakeholders in scientific inquiry and technological development. Broadening these conversations beyond academic and technical communities requires innovative approaches and inclusive spaces for dialog. In this context, the arts – an important reference for social knowledge and inclusion – can serve as a powerful medium to foster participatory, human-centered discussions around AI design and use.

The MoralPLai Project embraces this approach by combining scientific research with artistic exploration. Using research-based theater as a core method, the project investigates, communicates and reflects on current AI ethics research in a way that resonates with both expert and public audiences.

At the center of the project is an inquiry into how LLM-based chatbots shape ethical decision-making. Through expert interviews, the team examines their potential benefits and risks, broader societal implications, and the system requirements needed for their responsible development and use. These findings were not only published in academic journals but also transformed into a theatrical performance to engage wider audiences and promote AI literacy.

Namely, The Third Voice, the research-based play developed as part of the project, dramatizes the ethical and emotional tensions of using AI chatbots as moral advisors. It follows the story of a doctor and her daughter, both of whom consult a chatbot in a morally complex situation. The performance invites the public into a critical conversation about the role of AI in our lives, blending narration and research to spark reflection and dialog.

More than a performance, The Third Voice acts as a platform for reciprocal knowledge exchange: audience feedback and discussions are gathered to inform ongoing scientific research and contribute to the co-design of more human-centered AI systems.

Key research questions include:

• How can LLM-based chatbots support or impede human ethical decision-making?
• What system requirements are essential for their responsible design and use?
• How can research-based theater effectively engage diverse audiences in these pressing ethical questions?

The Third Voice premiered on May 22, 2025 at the Amerikahaus in Munich, followed by a panel discussion and open Q&A session to deepen engagement with the themes of the play.

Ultimately, this project serves as a proof of concept for integrating artistic practice into teaching, science communication and co-design in AI ethics, thereby paving the way for similar projects in the future.

“How should we proceed?”

This question, spoken as the final line of The Third Voice, was not merely a closing remark; it was an invitation to reflect on the heart of our inquiry: how to develop and use AI chatbots responsibly. But the question also points to the future of the MoralPLai Project itself. With the successful premiere of The Third Voice, the MoralPLai team reached an important milestone.

The MoralPLai team welcomes opportunities for collaboration. If you are interested in hosting The Third Voice, whether through a script-reading workshop, a collaborative research initiative, or through other formats, please contact the Project Lead, Dr. Franziska Poszler (franziska.poszler@tum.de). The team also encourages inquiries from those seeking to learn more about the project and its ongoing work.

#MoralPLai

Funding:

The project team gratefully acknowledge the financial support from the Notre Dame-IBM Tech Ethics Lab, the Friedrich-Schiedel-Fellowship Program at the TUM Think Tank and the TUM School of Social Sciences and Technology, the TUM Global Incentive Fund, the Tuition Substitution Funds, the Klaus Tschira Foundation gGmbH, TUM Center for Culture and Arts (CCA) as well as the TUM Global Visiting Professor Program.

Research Output:

Research Brief “Conversations With and About AI Chatbots: Perspectives From a Cross-institutional Survey on LLMs, Decision-making, AI Ethics and Science Communication”

Our MoralPLai Project in a Short Explainer Video