SPEAKERS
 Rislana Abdulazeez Kanya |  Iram Tanvir |
DESCRIPTION
Generative AI models are inherently stochastic; they "play dice" with every output, meaning they can hallucinate facts, exhibit bias, or provide inconsistent advice. In an educational setting, where accuracy and fairness are paramount, this randomness poses a significant liability. This panel tackles the complex web of accountability when AI systems fail or cause harm in learning environments. If an AI tutor provides incorrect medical information to a nursing student, or a grading algorithm unfairly penalizes a student due to dialect bias, who is to blame?
The session will dissect the chain of responsibility involving developers, educational institutions, teachers, and the students themselves. It explores the "black box" problem of AI, where the decision-making process is opaque, making it difficult to assign fault. The panel will likely discuss legal and ethical frameworks for liability, proposing models for "algorithmic accountability." It will also address the psychological impact on students who trust AI implicitly. By examining case studies of AI failures in education, the panel aims to establish clear guidelines for risk management, ensuring that the "dice roll" of AI does not compromise the integrity of the educational experience or the safety of the learners.
This session is part of the UNU Macau AI Conference 2026.
To view the full program, please visit the program page.
To join the Conference, please visit the registration page.
AI × Education: AI for Learning, Learning for AI
Nov. 25-26, 2026 | Macau SAR, China
AI Conference 2026 | AI for SDGs - Global Youth AI Future Innovation Competition | AI × Education Workshops
