Artificial Intelligence (AI) has evolved from self-improving computers to a multifaceted technology impacting our daily lives. With methods ranging from Machine Learning (ML) to data mining and advanced analytics, AI has revolutionised society, ushering in an era that demands AI literacy and comprehensive education. In this blog, we report on a literature review that explored the transformative potential of AI education in K–12 settings.
Bringing AI education to K–12
Teaching AI equips individuals with essential knowledge and programming skills, empowering them to become active participants in an AI-driven future. AI education is no longer confined to universities; there is a growing demand to introduce it in K–12 settings. Early exposure to AI concepts prepares young minds for a future where AI will play a significant role and encourages ethical AI use. Integrating AI into K–12 curricula empowers students with vital skills and nurtures future innovators and leaders.
The need for a systematic literature review
As AI education in K–12 is still an emerging area, conducting an in-depth literature review of empirical studies of classroom interventions was essential. Such a review helps gain valuable insights, identify gaps in current practices, inform policies, and foster collaboration among educators and researchers. With this in mind, Saman Rizvi and the research teams at the Raspberry Pi Foundation and the Raspberry Pi Computing Education Research Centre at the University of Cambridge conducted a systematic review of scientific literature published between 2019 and 20221.
Adhering to PRISMA guidelines, the review assessed AI teaching research interventions for students aged 4 to 18 years. The careful selection process involved considering only empirical studies that were peer-reviewed and published within the designated timeframe. Studies focusing on teachers rather than students, those with a small number of participants, non-English publications, or purely theoretical works were excluded. Through this comprehensive search across major bibliographic databases, a wealth of relevant literature was gathered. The final pool of 28 selected papers was analysed using content analysis to uncover both observable data and underlying meanings.
Valuable insights to inform future policies and research practices in AI education
The analysis revealed diverse pedagogical approaches and the use of interactive platforms and AI literacy games to engage students. Conceptual coverage of AI learning interventions was assessed across four levels: Social/Ethical, Application, Model, and Engine. Most interventions focused on the Model and Application levels, while attention to socio-ethical aspects grew over time.
- Measuring Students’ Success: The research evaluated both the affective and cognitive outcomes of AI education. Affective outcomes encompassed students’ attitudes, perceptions, motivation, and enthusiasm, while cognitive outcomes measured knowledge gain and skill development. Encouragingly, the results indicated a positive trend in affective outcomes, reflecting increased student enthusiasm and motivation.
- Topics and tools galore: AI education interventions covered a wide array of topics, ranging from fundamental supervised and unsupervised machine learning to advanced concepts like deep generative algorithms and conversational AI. Interactive tools like Scratch, gamification, and role-playing games played a pivotal role in engaging young learners.
- Teachers’ or educators’ involvement: Teacher involvement in AI education varied, with some studies being mainly researcher-led and others incorporating teachers as facilitators or co-designers. Engaging teachers in the research process proved to be empowering and beneficial for AI education. Their valuable insights and preferences, such as experience in using visual programming environments like Scratch, can significantly influence students’ AI learning experiences.
- Positive impact, yet room for growth: The review highlighted the positive impact of AI education in K–12 settings, igniting students’ curiosity and enthusiasm for AI concepts. However, it also identified areas for improvement, such as standardised evaluation frameworks and greater emphasis on socio-ethical considerations in future research and curriculum development.
Overall, the review highlighted the positive impact of AI education in K–12 settings, with students showing increased enthusiasm and motivation to explore AI concepts. However, the lack of standardisation and limited attention to socio-ethical aspects were identified as areas for further improvement in future research and curriculum development.
Key areas and future directions:
This systematic review has highlighted several crucial aspects that warrant deeper exploration. Let’s take a closer look at the key challenges and areas that require further attention:
- Bridging the divide: Equitable access to technology emerged as a pressing challenge. Moreover, students’ prior programming skills significantly impact AI learning outcomes. To ensure fair opportunities for all K–12 students in AI education, we must address the digital divide by providing the necessary infrastructure, devices, and internet connectivity in schools. Gender disparities in AI education were also evident, with older female students leaning towards non-technical aspects, and their male counterparts expressing more interest in coding and AI model training. Fostering diversity and inclusion is essential to promote equity and dispel biases in AI education.
- Inconsistencies and abstraction in AI topics: The review identified inconclusive findings across various AI topics taught in reported interventions, with disparities among different age groups and socioeconomic backgrounds. Resolving these discrepancies requires further in-depth research. Striking the right balance between abstraction and understanding is key to enhancing children’s interest in AI. Concealing certain complexities while teaching AI concepts can lead to better overall comprehension.
- Collaborative endeavour: Active engagement of educators in AI curriculum design and implementation is vital. The synergy among educators, researchers, policymakers, and industry professionals is key to fostering AI education in K–12 settings. This collaborative spirit and sharing of best practices can fuel the development of effective AI curricula. This collaboration will also ensure that AI education aligns with industry trends and demands.
Takeaway for researchers and practitioners
Addressing challenges like equitable technology access, gender parity, and educators’ involvement in curriculum design stands as a top priority for advancing AI education. Focussing on longitudinal studies and investigating how AI education influences students’ subject or career choices can inform future policy-making and curriculum development. Exploring diverse delivery methods, and enhancing conceptual coverage will also enrich the impact of AI education. Embracing interdisciplinary studies and translating findings into actionable policies will drive progress in this ever-evolving field. As AI education continues to evolve, we must remain open to refinement and improvement, harnessing the collective wisdom to prepare our students for an AI-powered world with knowledge, skills, and ethical awareness.
The literature review has been published in the journal “Computers & Education: Artificial Intelligence“, and it is readily accessible through this link.
- Rizvi, S., Waite, J., & Sentance, S. (2023). Artificial Intelligence teaching and learning in K-12 from 2019 to 2022: A systematic literature review. Computers and Education: Artificial Intelligence, 100145. ↩︎