Avoiding ‘Sinking the Boat’ While not ‘Missing the Boat’: K-12 Leaders' Early-on Perspectives of AI Risks and Benefits and Their Implications for Developers

This full research to practice paper reports on early perspectives from K-12 leaders regarding AI use in schools. With the advent of generative AI applications, K-12 leaders have a key role in providing (or precluding) access to students and teachers' uses of AI tools - as superintendents, principals and other district and school-level administrators will continue to be making (or at the very least informing) decisions about what AI tools will be made available as well as policies governing their use. These decisions will be informed by what K-12 leaders perceive are the potential risks and benefits of AI - as a key charge for K-12 leaders considering any innovation is to evaluate its potential to support student learning while reducing potentially harmful consequences. It is important to understand these current perceptions, especially for anyone designing applications of AI for K-12 education. While previous work has reported on teachers' views of AI, the perspectives of K-12 leaders, who serve as thought-leaders and decision makers, remain largely unexplored. U sing a semi-structured interview protocol, in late 2023, researchers interviewed 36 K-12 leaders across 23 districts in western New York state in order to gather their early perspectives regarding AI and to answer the research question: How do K-12 leaders perceive the risks and opportunities associated with using artificial intelligence in their school environments? Participants included superintendents, principals and various district and school-level administrators as well as some teacher leaders. These K-12 leaders articulated risks that can be categorized by four themes: (a) concerns regarding the ethical use of AI by both students and teachers (including cheating), (b) concerns around privacy and cybersecurity, (c) concerns around the accuracy or legitimacy of the output from AI systems and (d) concerns about replacing people/jobs. At the same time, these K-12 leaders recognized several important opportunities presented by AI, which should also be taken into consideration when making decisions, including (a) preparing students for the future, (b) improving potential for learning and instructional development and (c) supporting K-12 educators. Collectively, these risks and opportunities can be characterized with the idea that K-12 leaders were aware of the need to balance the risks in order to not “sink the boat” while also using care to not delay actions and potentially “miss the boat,” and which represents a more nuanced view of risk, consistent with what has been identified in the entrepreneurship literature. This work has implications for deliberate and informed decision-making regarding policies for and use of AI in the K-12 domain, and the supports needed for their adoption and effective use. The findings also provide valuable insights for developers of domain specific AI systems for K-12 schools. As computer scientists and engineers continue to train models, develop and select algorithms to serve schools, learners and educators, considering the risks and opportunities articulated by K-12 thought leaders and decision makers can support their work in advancing the technologies and potentially improving adoption.