Co-designing lesson plans for supporting human-centred AI literacy in K-12 education

The fast development of artificial intelligence (AI) has resulted in governments as well as national and international organizations calling for “AI for all", highlighting the need for AI-related competences (Unesco 2022). At the same time human centered approaches to AI are underscored, to protect human rights, promote sustainable development, and equip individuals with the necessary values and skills for effective human machine collaboration in everyday life and work (Unesco 2019). Still, AI in education remains a nascent field, with no clear guidelines on what to teach or how to pedagogically leverage AI (Luckin et al. 2022). Introducing AI competences, or AI literacy (Long & Magerko 2020), in K-12 education poses additional challenges as the topic is unfamiliar to most teachers and seldom has a clear space in the curriculum.

In this paper, we describe a human-centered process of co-designing (Steen 2013) AI lesson plans for grades 3-6 (students aged 9-13) together with students, teachers, IT educators and teacher trainers. By discussing the collaborative efforts to develop educational frameworks that enhance AI literacy among K-12 students, we emphasize the importance of integrating human-centered AI (HCAI) principles into educational practices (Shneiderman 2022). Thus, we ensure that students are given opportunities to both learn to understand AI technologies and engage with them in a way that respects human rights, values, and promotes sustainability.

We present the co-design phases involved, and lessons learned when creating educational experiences that 1) are suitable from both a teacher and student perspective, 2) can fit learning objectives in the current curriculum, and 3) prioritize the ethical, social, and human dimensions of AI. We outline the objectives of the co-designing project, which involved 46 Finland-Swedish stakeholders, including students, teachers, IT educators, and teacher trainers. This diverse group collaborated with us as researchers to create lesson plans that are not only informative but also engaging and relevant to students' lives. We adhere to a participatory approach (Nyholm 2022), where all stakeholders contribute their insights and expertise, ensuring that the resulting educational materials are well-rounded and effective. We outline the activities carried out in four co-design stages: co-ideation, co-design, co-implementation, and co-evaluation, including workshops and lessons held in grades 3 and 6 in two Finnish schools. We also describe the close to finalized lesson plans that may be implemented in AI literacy programs in schools.

While we are still in the process of analyzing the rich data collected during the co-design process, some preliminary findings can already be shared. One of the key themes found in the process is the importance of addressing biases in AI systems. We emphasize that as students explore how AI models can be trained, they must also understand the implications of the training data used, including issues related to representation and fairness. This aspect is reflected in two of the three lesson plans aiming to cultivate critical thinking skills, enabling students to question and analyze the technologies they encounter.

Another theme is the challenges faced by educators in integrating AI literacy in their teaching. These include a lack of resources, insufficient training, and the need for curricular guidelines that are adaptable to different educational contexts. The basis for the co-creation activities of the first lesson plan focused on providing teachers with sufficient resources, knowledge, and skills to present and discuss AI concepts in their classrooms. Consequently, the first lesson starts with discussions on what AI is and concrete examples of where AI can be found.

Lastly, we discuss the benefits and drawbacks identified during the collective effort of developing learning resources according to the co-design method (Viklund et al. 2023). Key benefits include mutual learning, the value of field observations and including multiple voices in the design process. Drawbacks include questions related to generalizability of the developed material, as the process makes a few teachers proxy voices for their colleagues, and the novelty effect of unfamiliar researchers holding lessons in the classrooms.

To conclude, we present a comprehensive overview of the co-designing process for AI literacy lesson plans in K-12 education. We underscore the necessity of integrating human-centered approaches into educational practices through collaboration among stakeholders. The insights provided in this paper serve as a valuable resource for educators, policymakers, and researchers interested in advancing AI literacy in schools. The co-design process and results highlighted in this paper are not relevant only in a Finnish context but also carry a wider international significance, as the results are language agnostic. We expect that the co-designed lesson plans and suggestions for progression will be applicable in both Nordic and international contexts.