WIP: Impact of Generative AI on Learning: Foci and Parameters for an Assessment Tool

This work in progress, innovative practice paper proposes a Learning and Teaching Impact Assessment Tool (TALIAT) that would help in assessing the sways of Generative Artificial Intelligence (GAI) on important aspects of Learning and Teaching (L&T). The focus of TALIAT is to identify the most relevant elements of GAI grammar and establish how they would reflect on Learning and Teaching practices. GAIsupported systems are generally perceived beneficial as they ensure resource availability, quick and ongoing feedback, higher level of interactivity, and continued engagement. GAI tools also inculcate the ability to perform independent research through progressive and systematic probing and consequent delivery of information. Nonetheless, issues pertaining to trust, liability, transparency, honesty, fairness in judgement and trustworthy reporting of student success need to be addressed for proper adoption of GAI applications. It is being acknowledged now that various stakeholders like system developers, administrators, policymakers, psychologists, educators, sociologists, and legal experts need to work together and agree upon standards and frameworks for producing professionally, ethically, and legally acceptable GAI systems. For these stakeholders to work together, an easy to follow and adopt assessment tool needs to be designed. This would facilitate establishing methods, and processes for confidently determining impacts of GAI technologies on Learning and Teaching. The current literature has yet to provide such facilitation. Such enablement would: improve teacher's perception of, and users' confidence in, GAI systems as tools like TALIAT would assist in assessing impacts of GAI tools prior to their deployment. The proposed TALIAT could also safeguard quality of information, student competence reporting, academic culture, and social and ethical values. TALIAT projects sixteen elements of grammar on seven important aspects of GAI implications namely, handling biases, prompt tuning and evaluation, output fairness, regulatory compliance, error detection, accountability and explainability. These seven aspects are deemed important for investigating the purpose, impetuses, scope, and L&T impact mitigation strategies of GAI applications. A set of thirty-two questions is included in TALIAT for GAI tools' impact assessment. TALIAT will be useful for the GAI stakeholders to work together as it enables informed decision making about adopting GAI systems and places rail guards for development of trustworthy GAI systems.