AI Literacy: Explaining Global Inequalities with ICILS 2023 and Policy Indicators

• In Event: Roundtable Session 9
  In Refereed Roundtable Session: Digital Learning, Global Competencies, and Equity: Innovations, Evidence, and Policy Perspectives

Proposal

The growing importance of artificial intelligence (AI) in civic engagement, employment, and education raises concerns about how equipped students are to interact with AI in fair ways. The extent to which present competences serve as the foundation for AI literacy readiness is not well understood, though worldwide surveys have shown persisting cross-national inequalities in digital skills (Fraillon et al., 2020). By examining the International Computer and Information Literacy Study (ICILS 2023), which assesses students' computational thinking (CT) and computer and information literacy (CIL). We see these concepts—which include the capacity to use algorithmic reasoning, navigate digital settings, and critically assess information—as crucial preconditions for AI literacy (Long & Magerko, 2020; Ng et al., 2021).

Four research questions are investigated: (1) What are the differences in CIL and CT results between OECD and non-OECD systems? (2) How well do teacher preparation, ICT availability, and socioeconomic background account for these disparities? (3) Are national AI curriculum and policies in line with students' preparedness? (4) In what ways do socioeconomic gradients change depending on the policy context?

To guarantee accurate estimations, the study uses jackknife repeated replication and plausible value approach (Rutkowski et al., 2010). By placing students in schools and nations, multilevel regression models will make it possible to estimate inequities within and between systems. Socioeconomic position, gender, ICT self-efficacy, and the frequency of ICT use at home and at school are all predictors at the student level. ICT resources and aggregated teacher indices for ICT confidence and pedagogical use are examples of predictors at the school level. We combine external policy data from the OECD AI Policy Observatory and UNESCO at the national level, accounting for teacher training programs, curricular integration, and the existence of AI strategies.

There are two ways in which this design goes beyond descriptive league tables. First, by emphasising the contributions of socioeconomic mix, school resources, and teacher preparation, it explains cross-national disparities rather than just reporting them. Secondly, it incorporates policy factors to evaluate if official pledges to AI and digital learning are associated with greater student preparedness.

The study makes three contributions. In concept, it reinterprets digital and computational abilities as the cornerstone of AI literacy preparation, offering a forward-looking perspective for analysing ICILS findings. It provides the first cross-national empirical test of how curriculum integration and the adoption of AI policies may either improve or worsen student results. From an analytical standpoint, it places the AI literacy gap both within and between OECD and non-OECD systems, where socioeconomic disparities could be more pronounced than those across nations (van Dijk, 2020).

Through the identification of equity gaps that run the risk of solidifying in the global knowledge economy and the clarification of whether nations investing in AI education initiatives are witnessing commensurate gains in student skills, the findings will inform both academia and policy. By doing this, the study offers much-needed proof for creating inclusive, cutting-edge educational systems that can equip all students for a world changing by artificial intelligence.

Author

• Liping Yang, University of Miami