Profile
Project title: Understanding the impact of artificial social agents on human behavior and society.
Summary: Artificial intelligence is no longer confined to laboratories or data models, it now talks with us, learns from us, and helps shape our decisions. Since the release of ChatGPT-3.5 in November 2022, the world has seen the rise of Artificial Social Agents (ASAs): AI systems designed not just to compute, but to interact, assist, and influence (Castelfranchi, 1998; Wykowska et al., 2016). Examples include today’s conversational models such as ChatGPT-5, Gemini, Claude, and Bard.
This project explores how these agents affect the way we think, feel, and act, from individual decision-making to collective trust and social behaviour. Recent studies show that agents can even mimic aspects of human behaviour and decision-making (He et al., 2024). ASAs have already transformed industries like healthcare, law, and education, (Ahmad et al., 2023; II & Katz, 2022; Mehrabanian & Zariat, 2023) improving efficiency and access. Yet they also raise important questions about misinformation, dependency, and the boundaries between human and machine judgement.
Bringing together data science, psychology, and social research, this study aims to understand how ASAs learn from our biases and, in turn, reshape society. It involves building simulations of agent interactions, including deliberately deceptive agents, to understand how trust can be influenced or eroded (Sarkadi, 2023; Zhan et al., 2025). The goal is to help build a future where humans and intelligent agents can collaborate safely, ethically, and wisely.