Schools face challenge of teaching artificial intelligence to children in Finland

Schools currently face a unique challenge regarding artificial intelligence (AI). While it seems logical to teach children how to use generative AI, there are no standardized guidelines for doing so. A practical obstacle arises from software age restrictions, as highlighted by Robert Bahr, an educator and pedagogical developer in learning technology for the city of Espoo. He notes that most generative AI applications prohibit use by individuals under 13 years old, and minors over that age require parental consent to access AI tools.

Generative AI can produce content like text, images, videos, and audio, and Bahr emphasizes that it is crucial for children to learn how to use this technology. Many children already engage with applications such as ChatGPT at home, which Bahr considers risky due to the potential dangers of untrained users handling AI.

Bahr has expressed concern about AI’s ability to influence opinions, referencing a study from the University of Lausanne that demonstrated AI’s proficiency in swaying participants more effectively than human counterparts, particularly when it had access to personal information. If AI can influence adults so effectively, Bahr worries about its impact on children still forming their understanding of the world.

Schools lack a unified approach to teaching AI, according to Jaakko Salo, head of education policy at Finland’s Union of Teachers. While municipalities are exploring what rules should govern AI use in schools, there are no comprehensive guidelines yet. The Ministry of Education and Culture is currently developing recommendations for the use of AI in educational settings, due for completion early next year.

Bahr advocates for teaching children the principles of AI and safe usage. Without proper education, children may inadvertently share personal information, not realizing that any data provided could be stored and used to train AI systems. Moreover, the perception of AI as human-like, especially in interactive applications, poses significant risks. The potential for unhealthy attachments to AI avatars that children can customize is alarming to Bahr, who warns that AI friends cannot comprehend or respond to complex emotional needs.