Letters from the Future: Exploring Ethical Dilemmas in the Design of Social Agents

Title of the Paper

Letters from the Future: Exploring Ethical Dilemmas in the Design of Social Agents

Paper Information

• Subject Area: Human-Computer Interaction, Ethics in Social Agent Design
• Keywords: Social agents, Human-Computer Interaction, Robotics, Artificial Intelligence, Ethics, Speculative Design, Design Fiction, Futures Studies, Cultural Probes

Research Background and Issues

• Identified Problems or Challenges:
  • Emerging technologies often lead to unforeseen social and ethical challenges, such as bias in AI algorithms, privacy issues on social media, and negative impacts of smart devices on children.
  • Design and development teams typically focus on optimistic visions of technological futures, often overlooking potential ethical concerns.
• Significance:
  • As social agents (e.g., robots and AI) evolve, they may introduce profound ethical conflicts and societal impacts. If these issues are not addressed in a timely manner, they could lead to greater negative consequences for human society.
• Research Motivation and Related Work:
  • Proposing design methods to uncover ethical issues during the early design stages of technology and promote deeper reflection.
  • The authors draw on methods such as Discursive Design and Experiential Futures, applying Ethnographic Experiential Futures (EXF) to social agent design for the first time.

Proposed Solution

• Proposed Method:
  • Using the Ethnographic Experiential Futures (EXF) framework to design probes in the form of "letters from the future," constructing possible future scenarios to document and analyze participants' ethical reflections.
  • Implementation process includes the following key stages:
    1. Map: Collecting expert insights on imagined futures of social agents through interviews.
    2. Mediate: Translating the scenarios mentioned in the interviews into tangible design probes ("letters").
    3. Mount: Delivering the design probes to participants via mail.
    4. Map: Gathering and analyzing participants' feedback.
• Innovative Aspects:
  • The EXF method combines qualitative investigation of the future with design practice, not only collecting future imaginaries but also enabling participants to reflect through interactive formats.
  • Integrating remote, asynchronous experiential design with high-fidelity visualized letters, exploring a low-cost yet immersive mode of expression.
• Implementation Steps and Techniques:
  • The authors designed three specific future letter probes:
    1. Educational Agent: Depicting the role of personalized agents assisting students in learning and exploring ethical issues related to children's behavior.
    2. Value Agent: Tracking user behavior to evaluate and guide alignment with personal values.
    3. Privacy Agent: Designing an autonomous agent to combat data collection and surveillance on behalf of users.
  • Feedback on future scenarios was collected through the mailing of these letters.

Research Findings

• Specific Findings:
  • The study uncovered three underexplored ethical issues in the design of social agents:
    1. Whether data privacy and algorithmic ethics fall within the scope of Human-Robot Interaction (HRI) or Human-Agent Interaction (HAI).
    2. Whether social agents should collect users' social interaction data and, if so, to what extent.
    3. Whether ethical evaluations of technological development should be the responsibility of industry insiders or policymakers.
• Advantages:
  • Compared to traditional exhibit-based future probes (e.g., museum exhibitions), this method enables deeper documentation of participants' reflections.
  • The high-fidelity yet low-cost letter probes successfully engaged participants and fostered in-depth discussions.
• Experimental or Evaluation Results:
  • Participants not only expressed specific thoughts on the probes but also questioned the ethical boundaries of social agent design.
  • The temporal distance and narrative roles embedded in the letters influenced participants' acceptance of the scenarios.
• Limitations and Future Directions:
  • Feedback was collected via online questionnaires, which limited the depth of dialogue about the probes.
  • As the probes spanned multiple domains, the study did not deeply compare the ethical commonalities and differences across various future scenarios.
  • Future research could focus more intensively on a single design space for in-depth exploration or use group discussions to enhance the social and interactive nature of data collection.

In summary, this paper provides a framework for exploring ethical issues during the early stages of technological design through the EXF method and proposes a low-cost, efficient speculative design approach. However, there remains room for improvement in the methods of data collection and the comparability of probe designs.