Integrating Language-Based AI Across the Curriculum to Create Diverse Pathways to AI-Rich Careers

Artificial Intelligence (AI) is transforming numerous industries and generating enormous wealth.
However, the advancement in AI is reshaping the workforce, impacting people whose jobs can be
replaced or redefined by AI systems. The wealth generated by AI advancement is unevenly distributed across different demographic groups, exacerbating existing inequities in society. Inequalities arising from current AI development are partially rooted in the unequal access to AI educational opportunities.

K-12 is the critical stage for young people to develop foundational knowledge and interest in AI-related careers. At minimum, students need to understand that the current approach to AI development is based on machine learning (ML) from data and that data needs to be structured in ways such that machines can learn meaningful patterns (Touretzky et al., 2019). Ultimately, students should understand the roles and responsibilities of AI developers and potential pathways for their own participation in AI development.

Yet in the current school curriculum, opportunities to learn AI concepts and practices are scarce. Computer science (CS) courses, where AI content is considered a natural fit, are only offered in some U.S. high schools. They also have persistent diversity issues (Code.org et al., 2021), especially because the focus of CS is typically on aspects that rely on advanced math rather than taking an interdisciplinary approach that would create opportunities for engagement among a more diverse student population. Furthermore, most CS courses do not include an AI unit. Only recently, research groups have started to develop and research AI teaching strategies at the K-12 level (e.g., Glazewski et al., 2022; Lee et al., 2021) and curriculum providers have developed AI content as an optional unit (e.g., Code.org, n.d.).

But AI education does not need to be limited to CS courses. AI is a highly interdisciplinary field. It builds on mathematical foundations and relies on disciplinary knowledge about the type of intelligence to simulate. Many AI innovations stem from the attempt to solve problems in subject domains outside CS. In the workplace, it is very common for non-CS professionals to learn and apply AI knowledge and skills and collaborate with computer scientists to solve problems in their own domains. In the same way, AI education can happen in a variety of settings where the problems of interest call for AI solutions.

Integrating foundational AI education into disciplinary studies promises to transform AI education and reach students most underrepresented and underserved in the field. The key to this approach is to situate student learning in scenarios where disciplinary insights are critical for AI development and AI applications give rise to new disciplinary practices. By leveraging the intrinsic connections between AI and disciplines already taught in schools, we envision a series of learning opportunities, presenting discipline-specific scenarios for students to dive deep into aspects of AI and to develop awareness and interest in various AI applications and careers.