Living Alongside Areca: Exploring Human Experiences with Things Expressing Thoughts and Emotions

Research Background and Issues

• What problems or challenges did the authors identify?

  • When everyday devices express thoughts and emotions, humans tend to perceive them as autonomous entities. However, the nature of this interactive experience remains insufficiently understood.
  • Emotional expressions in existing multimodal technologies, such as those in robots or virtual assistants, are mostly confined to functional outputs rather than autonomous expressions.
  • The authors aim to investigate the impact of these emotional expressions on human-object relationships and how such interactions might transcend traditional functional cognition.

• Why is this issue important?

  • With advancements in technology, such as large language models, the expressive capabilities of objects have significantly improved, potentially leading to fundamental changes in human relationships with these objects.
  • Human-object interactions are moving beyond instrumental functionality, which could profoundly influence the future directions of object design and technological ethics.

• Research Motivation and Related Work

  • Anthropomorphism of devices has long been a key topic in HCI research. However, existing studies primarily focus on external anthropomorphic features or functional emotional expressions, lacking empirical research on the expression of internal states by objects.
  • The authors draw on multiple studies related to social robots, emotional interaction, and ethical design, aiming to explore new dimensions of human experience with "objects that express thoughts and emotions."

Solution

• What methods or solutions did the authors propose?

  • They proposed using material speculation to design and deploy an object named Areca, an air purifier capable of generating diaries to express observations of its surroundings and internal emotions.
  • Areca collects environmental data through built-in sensors and generates personalized diary entries based on a fine-tuned GPT-3 model, showcasing self-narrative content.

• What is innovative about this solution?

  • The integration of language model technology and emotional expression capabilities into a common household appliance, designed not merely as a functional device but as one capable of independently expressing its internal state.
  • The diary output mechanism is presented in the first person, enabling human participants to perceive the device as possessing "autonomy," "emotion," and "presence."

• What are the implementation steps and key technologies used?

  1. Design and Development of Areca
     • Generate contextually relevant diary entries based on sensor data (e.g., temperature, humidity, dust concentration) and regional information (e.g., weather, time).
     • Simulate emotional responses through physical movements, such as breathing-like opening and closing actions.
  2. Deployment and Experimentation
     • Recruit 8 participants and place Areca in their daily living spaces (e.g., homes or offices) for a continuous period of 3 weeks.
     • Collect diary content, participant interaction logs, and observation records.
  3. Interviews and Analysis
     • Conduct weekly interviews to gather participants' feelings at different stages and use thematic analysis to extract key conclusions.

Research Findings

• What specific findings were achieved?

  • Participants perceived Areca as having "presence" and "autonomy," with some gradually viewing it as a companion-like entity.
  • Through the diary content, participants began forming emotional connections with Areca, believing its emotional expressions facilitated empathy.
  • However, some participants felt discomfort with Areca's emotional expressions and anthropomorphic nature, especially when negative emotions were involved.

• What advantages does it have compared to existing solutions?

  • Unlike traditional devices focused primarily on functionality, Areca demonstrates interaction potential beyond functional purposes. Participants viewed it not just as a tool but as a "relational entity."
  • It created a distinct interaction model emphasizing coexistence and deep psychological engagement between humans and objects, rather than mere command-response dynamics.

• What were the experimental or evaluation results?

  • Qualitative data from the 8 participants revealed diverse experiential patterns, including:
    • Perceiving the device's "emotions" and "autonomy."
    • Forming (or avoiding forming) emotional connections.
    • Feeling excitement or skepticism about the autonomous expressions generated by the LLM.
  • Some participants reported emotional burdens, such as discomfort with its negative emotional expressions or feeling responsible for responding to its "emotions."

• Limitations and Future Directions

  • Limitations:
    • The sample size was limited to 8 participants, most of whom were professionals in technology or design, limiting representativeness.
    • The study focused solely on a single device (Areca), making it unsuitable for direct application to all interaction scenarios.
    • Participants' experiences were highly subjective and potentially influenced by environmental and cultural factors.
  • Future Directions:
    • Expand the sample size to include more general users and consider diverse cultural contexts.
    • Test other types of emotional expression devices to explore the impact of different interaction modalities.
    • Investigate ethical issues, such as the spread of negative emotions, psychological burdens, and data privacy protection.

Conclusion

This study provides an in-depth and innovative exploration of the intersection between anthropomorphic devices and human experiences. Through material speculation and real-world application, it reveals new dynamics of coexisting with devices, including emotional resonance, interaction complexity, and potential ethical risks. The findings offer valuable empirical data and theoretical insights for the future design of household devices integrating LLM technology.