Designing a Dashboard for Transparency and Control of Conversational AI

TL;DR

This paper presents a transparent conversational AI dashboard that displays and allows control over user-related internal states, improving user understanding, exposing biases, and enhancing perceived control.

Contribution

It introduces a novel dashboard integrating interpretability techniques with user experience design to make chatbots more transparent and controllable.

Findings

Users appreciated seeing internal states and found it increased their sense of control.

The system exposed biases in the language model's behavior.

Participants provided valuable feedback for future improvements.

Abstract

Conversational LLMs function as black box systems, leaving users guessing about why they see the output they do. This lack of transparency is potentially problematic, especially given concerns around bias and truthfulness. To address this issue, we present an end-to-end prototype-connecting interpretability techniques with user experience design-that seeks to make chatbots more transparent. We begin by showing evidence that a prominent open-source LLM has a "user model": examining the internal state of the system, we can extract data related to a user's age, gender, educational level, and socioeconomic status. Next, we describe the design of a dashboard that accompanies the chatbot interface, displaying this user model in real time. The dashboard can also be used to control the user model and the system's behavior. Finally, we discuss a study in which users conversed with the instrumented system. Our results suggest that users appreciate seeing internal states, which helped them expose biased behavior and increased their sense of control. Participants also made valuable suggestions that point to future directions for both design and machine learning research. The project page and video demo of our TalkTuner system are available at https://bit.ly/talktuner-project-page