Generative AI in education: Current state, future directions, and implications in Bangladesh

• In Event: The role of generative artificial intelligence (AI) in Education: inclusion, innovation, and impact

Proposal

1. Introduction

Generative AI holds immense promise in the education sector, particularly by enhancing personalized learning and broadening access to education, in line with SDG 4 – Education 2030. However, concerns about its misuse, such as promoting cheating and plagiarism, have surfaced (UNESCO, 2023). While generative AI is advancing rapidly worldwide, Bangladesh is lagging in its adoption due to an expanding digital divide, worsened by the COVID-19 pandemic. The IMF AI Preparedness Index reflects this disparity, with Bangladesh scoring significantly lower (0.38) compared to developed countries (over 0.7) and even trailing other South Asian nations like India (0.49) and Sri Lanka (0.44) (IMF, 2023).

In a developing country like Bangladesh, marked by high inequality, generative AI could further deepen this digital divide (Hendawy, 2024). Factors such as digital literacy, cultural norms, and economic conditions influence access, leaving individuals unfamiliar with technology struggling to use AI tools effectively. Additionally, recent political instability may have further weakened Bangladesh’s technological sector, underscoring the need to assess the nation’s current standing in AI adoption (Wang et al., 2024).

This study aims to evaluate the knowledge, capacity, and attitudes of Bangladeshi educators, learners, and professionals toward generative AI, investigating the barriers to access and its potential consequences. Furthermore, it seeks to gather insights on the need for policy regulations governing AI. The findings will offer crucial recommendations for policymakers, educational administrators, government officials, researchers, and AI developers looking to promote the adoption and regulation of generative AI in developing countries like Bangladesh.

2. Theoretical Framework

The study utilizes the Theory of Planned Behavior (TPB) as its theoretical framework to predict attitudes, subjective norms, and perceived behavioral control towards generative AI in education, which in turn forecast its adoption and integration (Ajzen, 1991).

2.1. Attitude Towards Generative AI

Attitude is shaped by perceived knowledge and capacity (Fabrigar et al., 2006), perceived usefulness and ease of use (Davis et al., 1989), optimism (Conversano et al., 2010), and AI anxiety (Cho & Seo, 2024).

2.2. Subjective Norms

This refers to the social expectations influencing the adoption of AI, which are affected by the support from institutions and peers (Ajzen, 1991; Omer et al., 2015; Peters & Templin, 2010).

2.3. Perceived Behavioral Control

This reflects the perceived ease or difficulty of using AI, influenced by barriers, challenges, and consequences, which can either decrease or enhance individuals' confidence and intention to adopt AI (Ajzen, 1991).

The attitude, subjective norms and perceived behavioral controls will act as independent variables, while future behavior or use of generative AI will be the dependent variable. The independent variables will be explored based on the above-mentioned indicators.

3. Methodology

The study will employ a quantitative research approach utilizing a survey design. Data will be gathered through a questionnaire that includes both closed and open-ended questions, targeting teachers, learners, and other education professionals. Learners will be defined as students from secondary to tertiary education levels. Simple random sampling will be used for selecting respondents, with the questionnaire distributed online via the Institute of Informatics and Development’s website. Additionally, convenient sampling will be employed by mailing the questionnaire to education professionals within the organization’s network.

4. Analysis

Descriptive statistics will summarize the key variables in the study. Correlational analysis will examine potential digital divides in access to or use of generative AI based on socio-economic conditions or gender. Additionally, regression analysis will be performed on indicators of attitude, subjective norms, and perceived behavioral controls to forecast future AI usage in education. The primary data analysis tools will be IBM SPSS version 25 and, if needed, IBM AMOS version 25.

5. Expected Findings

It is expected that there will be varied levels of knowledge and attitudes—both positive and negative—towards generative AI among Bangladeshi educators and learners. It is likely to uncover significant barriers to AI adoption, such as limited resources, inadequate training, and infrastructural constraints. Additionally, the findings are expected to demonstrate that high perceived barriers and negative consequences lower perceived behavioral control, thereby reducing the intention to engage with AI technologies. The research is also expected to emphasize the necessity for targeted policy regulations and support mechanisms to enable effective integration of AI in the education sector.

6. Conclusion

With the rapid advancement of generative AI, the digital divide is widening at an exponential rate. Therefore, it is crucial to understand the factors shaping the current state, future directions, and implications of AI in Bangladesh’s education sector.

Authors

• Maisha Tasnim, Institute of Informatics and Development (IID)

• Aruja Islam, Institute of Informatics and Development (IID)