As social robots are increasingly endowed with human-like personas, their design and behavior raise critical questions about deception, ethics, and trust in Human-Robot Interaction (HRI). The exploration of both the intentional and unintentional deception embedded in social robots’ persona design is necessary as generative artificial intelligence (GenAI) and increasingly anthropomorphic robot appearances shape conversational capabilities. While WoZ-controlled robots rely on pre-scripted responses, LLM-integrated robots generate real-time responses, simulating an adaptive persona. However, concerns arise as human interlocutors misinterpret intent or overestimate capabilities despite disclosed limitations (e.g., LLM hallucinations, lack of mobility) due to anthropomorphic features like an expressive face or LED-lit 'brain'. Such cases highlight the risks of external, superficial, and hidden state deception in HRI, facilitated by robot persona, directly influencing user expectations and trust in social robotic systems. A key question is whether users recognize these personas as constructed identities, unknowingly accept the deception, differentiate robot autonomy between WoZ-controlled and LLM-integrated robots, or strategically engage with it for smoother interaction. This workshop will examine how inherently deceptive robot personas influence trust and ethical considerations in long-term interactions, aiming to establish a common consensus on responsible persona design in HRI. By catering to an interdisciplinary discussion, we strive to define transparent and trustworthy approaches for real-world social robots.

• Submission deadline: Mon 21 July 2025
• Author notification:  Mon 28 July 2025
• Workshop: Mon 25 August 2025

Please submit your short paper (4 pages) or poster in PDF format: Submission Link (use pass: llm_robo_id4)

Topics for participation include (but are not limited to):

• Voluntary and involuntary deception in HRI experiments
• The design of social robots that increase or decrease the risk for social deception and uncalibrated trust and expectations
• Measuring Deception in HRI – Methods and challenges in assessing when and how humans feel deceived by robots
• The theory and practice of robot persona design
• Ethical aspects using LLMs and WoZ paradigms in experimental studies
• Long-Term Effects of Deception in HRI on user trust, acceptance, and transparency

Keywords: human-robot interaction, social robots, robot persona, large language models, deception, ethics