Talking to Strangers
Richard Lachman · 2024
This chapter is a critical and ethical examination of the evolving landscape of artificial intelligence (AI) conversational interfaces, including chatbots as well as large language models (LLMs). The use of human-like modes of communication, specifically the fluency of natural language, can trigger false assumptions about the trustworthiness, reliability, transparency, and safety of AI systems. As these tools are increasingly used in domains such as healthcare, mental health support, education, and employment, it becomes increasingly important to correctly link user assumptions of quality and reliability to the capabilities of AI. Interfaces should be aligned with the capabilities of the underlying systems they govern and should avoid misleading design cues. Researchers and corporate developers must prioritize social responsibility, using transparent design elements and coordinating the signifiers of their interfaces with the reliability and capabilities of their AI tools.