UNICEF Guidance on AI for Children: Application to the Design of a Social Robot For and With Autistic Children
Project Overview
The document examines the integration of generative AI in education, particularly through a project that employed the social robot Pepper in a Special Educational Needs (SEN) school to enhance the well-being of autistic children. By adopting a participatory design approach, the project actively involved both autistic children and educators in refining the robot's functionalities, ensuring that it met the specific needs of its users. This initiative aligns with UNICEF's guidelines on AI for children, emphasizing the importance of fostering development, inclusion, fairness, and safety in educational settings. The outcomes of the study were notably positive, as the robot facilitated emotional engagement and learning opportunities for the children. However, it also highlighted challenges, such as the potential for excessive emotional attachment to the robot and the necessity for careful ethical considerations in its deployment. Overall, the findings underscore the promise of generative AI in creating supportive educational tools while also calling for a thoughtful approach to its implementation to safeguard the interests and well-being of children.
Key Applications
Social robot (Pepper) for supporting well-being of autistic children
Context: Special Educational Needs (SEN) school with autistic children aged 13-16
Implementation: Participatory design involving focus groups with children and teachers to co-create the robot's behaviors, followed by a three-week deployment in the school.
Outcomes: Improved emotional well-being and social support for autistic children; children expressed excitement and positive feelings about interacting with the robot.
Challenges: Some children expressed fear or dislike of the robot; emotional attachment raised concerns about the robot's removal impacting children.
Implementation Barriers
Ethical
The need to ensure the emotional and psychological safety of children while interacting with the robot.
Proposed Solutions: Training staff to recognize distress, ensuring the robot does not store personal data, and having an exit strategy for the robot's removal.
Technological
Technical limitations of speech recognition technology for non-verbal children.
Proposed Solutions: Designing hybrid interaction mechanisms where children can respond using icon-based buttons.
Project Team
Séverin Lemaignan
Researcher
Nigel Newbutt
Researcher
Louis Rice
Researcher
Joe Daly
Researcher
Vicky Charisi
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Séverin Lemaignan, Nigel Newbutt, Louis Rice, Joe Daly, Vicky Charisi
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai