Search

In this position paper, the authors lay out the imperative for equitable artificial intelligence (AI), highlighting the essential role of access-oriented institutions and calling on technology companies (both large and small), foundations, and local, state, and federal regulators to consult with the newly convened Complete College America Council on Equitable AI in Higher Education. Their belief is that equitable AI spans far beyond the risk of mis-trained data. How schools adopt or reject these tools, the priorities of AI vendors, access to resources that enable the use of these tools, and the systemic integration of historically underrepresented and underserved voices will shape whether technology amplifies privilege or fosters inclusivity. A three-fold framework is presented for understanding Equity in AI, considering not just the quality and unbiased nature of the data used to train generative AI machines but also who has access to conversations around policy and product, as well as which institutions have access to the resources and safety nets that enable innovation and experimentation in the field of AI. A disruptive new advisory council is proposed, the Complete College America Council on Equitable AI in Higher Education, composed of representatives from historically excluded institutions and, by extension, students. The authors urge policymakers, technologists, and funders to proactively consult the Council and disrupt systemic inequities by integrating AI into higher education rather than continue to perpetuate them. [This paper was created in partnership with T3 Advisory.]

To trace the evolution of research strands in published Computer Assisted Language Learning (CALL) papers over time, a corpus of 426 highly-cited papers in four major CALL journals ("ReCALL," "CALL," "LL&T," and "CALICO Journal") was compiled and coded using NVivo 12. All identified aspects of technology-mediated language learning and teaching were collected to produce a comprehensive list of 690 recurring research strands and then, by adopting a constant comparison method, were merged to form 119 unique research strands. The top 10 alone represent almost half of all research strands: Computer-Mediated Communication (CMC), writing, vocabulary, feedback, evaluation, learning environment, telecollaboration, design, speaking, and grammar. This paper provides the rationale for the study along with the methodology for data collection and analysis, with a particular focus on the dominant and intermediary strands to inform future CALL publications. [For the complete volume, "Intelligent CALL, Granular Systems and Learner Data: Short Papers from EUROCALL 2022 (30th, Reykjavik, Iceland, August 17-19, 2022)," see ED624779.]