CupidGPT or Love Bytes: A Turning Test in the World of GenAI Matchmaking

Supervisor: Dr Alan Woodley

With the increasing sophistication of Generative Artificial Intelligence (GenAI), large language models (LLMs) like OpenAI’s ChatGPT, have raised questions about their ability to replicate human-like social interactions. While prior research has explored GenAI’s applications in professional and empathetic domains, its role in romance remains largely unexplored. CupidGPT or Love Bytes: A Turing Test in the World of GenAI Matchmaking investigates whether GenAI can convincingly participate in a romantic matchmaking scenario. Inspired by the Turing Test and traditional dating shows, this study places a human participant in conversation with two suitors – one human and one AI – analysing whether AI can generate responses indistinguishable from a real person. This research explores AI’s capacity for emotional engagement, authenticity, and trust in interpersonal relationships. By reviewing existing literature on AI empathy, designing a web-based Dating Simulation and Turing Test, and statistically analysing participant responses, the study aims to assess the effectiveness of AI in replicating human emotional intelligence. Preliminary findings are expected to provide insights into the limitations and potential of GenAI in romantic contexts. By analysing human perceptions of AI-generated responses, this research contributes to broader discussions on AI-human interaction, digital companionship, and the evolving role of AI in social relationships. The outcomes will inform future developments in AI-driven matchmaking systems and raise ethical considerations regarding the emotional impact of AI on personal relationships.