Data literacy and artificial intelligence – biases and limitations

1. Introduction

One of the most significant challenges to the responsible use of AI in higher education is data bias, which can result from various factors, such as inadequate samples, collection and measurement errors, algorithm design limitations, or subjective interpretations reflected in the data. These biases can lead to unfair educational outcomes for students, undermine trust in AI technologies, and reduce their effectiveness in teaching and evaluation.

AI systems, though advanced, cannot independently grasp the complexity of educational environments. They rely on the quality and representativeness of the data provided, making human oversight and critical analysis of results essential. Teachers, researchers, IT professionals, and other educational stakeholders should work together to assess data sources, their interpretation, and the ethical implications of their use.

The following outlines the main types of data bias, their potential consequences in education, the limitations of AI systems, and recommendations for the ethical and responsible use of data in practice.