Textual Intelligence

7Misbehaving AI Models and AI Interaction Issues with Humans

Nishi Gupta¹* and Shikha Gupta²

¹Gurugram University, Gurugram, Haryana, India

²Maharaja Agrasen Institute of Technology, GGSIPU, Rohini, Delhi, India

Abstract

In our daily lives, Artificial Intelligence (AI) is becoming more and more common. Examples include self-driving cars and personalized social media suggestions. As AI systems become more complex, concerns about their misbehavior and the challenges they pose for human interaction are growing. This chapter provides an insight into the current literature on misbehaving AI models, their impact on human interactions, and ethical considerations. It discusses various types of misbehavior exhibited by AI models, including biased decision-making, adversarial attacks, and unintended consequences. It also examines the impact of these misbehaviors on human interactions, such as distrust, frustration, and disengagement. The chapter also explores the ethical considerations of AI misbehavior, including issues of accountability and responsibility. One key area for future research is the development of methods to detect and prevent AI misbehavior. Another important direction is the exploration of the psychological and sociological factors that influence human interactions with AI systems. Overall, this paper provides a comprehensive overview of the issues related to misbehaving AI models and their impact on human interaction. It highlights the need for continued research in ...