Project description

Since the introduction of ChatGPT in November 2022, the use of artificial intelligence (AI), especially AI-based chatbots, has become accessible to many people. This project investigates how people use ChatGPT and comparable chatbots for knowledge search and as a creativity tool, both in professional and private settings. The studies thus contribute to the practice fields knowledge work with digital media and internet use.

Many search engine providers integrate AI-based chatbots, which are based on so-called large language models (LLMs), into their services, meaning that they are also used by many for knowledge processes such as information searches. However, the adoption of AI-generated content harbors risks, as the technology also presents incorrect information very confidently.
The current research project aims to better understand the opportunities and risks of using LLM-based chatbots in everyday professional and private life. Two initial experiments show that the way in which information is presented by ChatGPT or comparable chatbots influences the credibility assessment. The same information is perceived as more credible when it is presented as a dialog - whether typed with ChatGPT or spoken by a voice assistant such as Alexa - than when it is presented as static text. This type of presentation also makes it more difficult to identify incorrect information. In survey studies with representative samples and a longitudinal study with knowledge workers, we also investigate how people use LLM-based chatbots for different tasks, to what extent this is associated with information benefits and higher creativity, and which factors predict successful use.

Cooperations

• Dr. Frank Schneider, University of Amsterdam (The Netherlands)