Skip to main content

Ethics in AI

Fundamental ethical principles​

The European Commission defines the following fundamental ethical principles of AI:

  • Human Action and Supervision: AI should act as a tool to support humans, not to replace their autonomy. Important final decisions must be made by people.
  • Technical Robustness and Security: Algorithms must be secure, reliable and robust enough to not cause unintentional damage.
  • Privacy and Data Governance: The privacy of student data must be guaranteed and clear protocols established on who has access to it and what it is used for.
  • Transparency: It must be possible to trace the operation of an AI system. Users should know that they are interacting with an AI and understand (to the extent possible) how it makes its "decisions."
  • Diversity, Non-Discrimination and Equity: AI systems must be accessible to all and must not introduce or amplify biases that harm certain groups of students.
  • Social and Environmental Wellbeing: AI should be used to promote positive social change, including quality education.
  • Accountability: Mechanisms must exist to ensure responsibility for AI systems and their results.
Biases as a consequence of how it works

Biases in AI are a direct consequence of its statistical nature: models reflect the regularities of the data on which they were trained. Understanding this mechanism is necessary for ethical use of these tools. The types of biases, their causes, and examples are discussed in detail in the [How language models work] section (11-functioning.mdx).

Special considerations for teachers​

As a teacher, you do not need to be an expert in machine learning, but you do need to be a critical and conscious user. Your role is essential to mitigate the risks of AI in the classroom. When selecting AI tools, you should evaluate three critical aspects: transparency of operation (avoiding unexplained "black boxes"), absence of bias through independent audits, and compliance with privacy regulations such as the GDPR, especially when dealing with highly sensitive educational data.

In everyday use, it is essential to not delegate critical judgment and use AI as an assistant, not an oracle, constantly monitoring recommendations to avoid pigeonholing students. The most important aspect is promoting AI literacy among students: teaching them how these tools work, promoting critical thinking towards their results, and establishing clear ethical standards for use. Finally, you must act as agent of equity, observing the differential impact on various groups of students, reporting on detected problems and guaranteeing equitable access to prevent AI from generating new digital divides.

Ethical integration of AI is a continuous process of learning and adaptation. Your role as a teacher is irreplaceable in guiding students in this new paradigm, ensuring that technology serves to enhance human learning within a framework of justice and equity.