Understanding Communicative AI in Science Communiation

Communicative AI (ComAI) refers to technologies that process and generate human language, carrying out communicative tasks that were traditionally performed by humans. These systems are increasingly used in journalism, education, and public engagement with science. However, their influence on the quality, accessibility, and trustworthiness of science communication remains an open question.

Wiss-KKI is a multi-institutional research project led by Prof. Dr. Helena Bilandzic (University of Augsburg), Prof. Dr. Monika Taddicken (Technical University of Braunschweig), and Prof. Dr. Björn Schuller (Technical University of Munich). The project brings together experts in science communication, artificial intelligence, and media studies to investigate how AI is shaping public discourse and engagement with science.

We adopt a three-phase approach:

• Conceptual Phase: Developing theoretical frameworks for evaluating science communication about and with ComAI.
• Empirical Phase: Conducting content analyses, experiments, and surveys to assess ComAI’s role in shaping public discourse and engagement.
• Technical Phase: Translating research findings into guidelines and tools for ComAI-assisted science communication.

Our goal is to critically examine how ComAI transforms science communication and to develop knowledge, frameworks, and tools that promote ethical, effective, and inclusive AI integration. We explore how ComAI is represented in public discourse, how users engage with ComAI-generated science content, and how it can be leveraged to improve communication while minimising risks such as misinformation and bias. Through interdisciplinary research, we aim to ensure that ComAI-driven science communication is transparent, responsible, and beneficial to society.

---