SARs are robots that interact with users through social engagement and support. Equipped with AI-enabled features, SARs are increasingly being adopted in homes, schools and hospitals to improve accessibility, monitoring and assistance. Their diverse functionality allows them to serve as companions or health monitoring tools for seniors, as well as to assist children with special needs in developing their social skills.
Deep learning, facial and voice recognition, and adaptive learning features allow SARs to foster long-term engagement with their users by making interactions feel more human-like, even if the robot itself is not humanoid. While SARs’ ability to remember user preferences, routines and personal history allows them to create more personalized user experiences, this functionality has also spurred discussion around ethical risks and the potential for malicious actors to abuse human robot relationships for emotional manipulation or deception.
The 2024 AI for Good Global Summit featured a Robots for Good workshop to foster discussion around ethical guidelines and standardization in the design and development of SARs. Robotic researchers emphasized the need for transparent decision-making and communication to convey what values are being embedded into robotic technology and the expected positive or negative outcomes from their deployment. Consideration of why a robotic solution is required in the first place, identification of whom the robot affects, and a cost-benefit analysis of the immediate and long-term impacts of its implementation are core determinants in guiding thoughtful deployment. Further, the capabilities and limitations of a robot should be communicated clearly to ensure that its potential risks are not obscured by the allure of its benefits.
The findings from this workshop highlight the need for a systematic approach to designing and developing SARs that considers their potential impact on individuals and society as a whole. This effort includes the development of standardized metrics for assessing “good” outcomes across different countries and further contextual, local or temporal contexts. Additionally, the design and assessment process requires active community involvement that captures multidisciplinary and diverse perspectives to consider best practices.
