His district offers Cyber Week, an optional week during the summer for teachers to explore innovative teaching practices. This past summer, Cyber Week’s topic was AI.

Additionally, the district has low stakes, monthly, hour-long meetings where teachers can explore generative AI without the expectation that they incorporate it immediately into their classroom or teaching. “I believe that the lack of that expectation of an outcome…breeds more innovation in our schools,” said Guidotti.

AI’s Broader Implications

Part of experimenting with AI is about helping teachers enhance instruction.

Although AI tools might seem helpful in day-to-day tasks, generative AI tools for instructional design must have a critical lens, according to Marc Watkins, director of the Mississippi Institute for Teachers.

Watkins pointed to Harvard’s AI Pedagogy Project as a resource for teachers looking to learn more about the ethical use of AI and practical tools. The Modern Language Association has also collaborated with Conference on College Composition and Communication to form a task force on writing and AI dedicated to developing guidelines and resources.

“When it comes down to creating content that’s for distribution for students, we ask the teachers to be transparent” about the use of AI generated activities or lesson plans, said Guidotti. When a teacher discloses their own use of AI to students, it creates an opportunity for a broader conversation about when it may or may not be appropriate to use AI in an education setting, he added.

According to Dukes, AI isn’t particularly good at creating curricula. Instead, he suggested using AI to generate creative word problems and activities that fit into an existing curriculum.

“Experimentation [with AI] can be useful and fun, particularly if the teacher is intellectually engaged in that process and is paying close attention because AI makes a lot of mistakes,” said Dukes.

Dukes also warned about explicit and implicit biases when it comes to using tools like AI detection software and AI grading especially if the output will be assessed for punishment or disciplinary actions. “[Teacher] biases are going to shape the decisions they make about who to investigate, and that has implications,” said Dukes.

Protecting student privacy and data, copyright infringement, and disclosure of use are also big ethical implications to consider when using AI as a teacher. For example, “you definitely don’t want to give ChatGPT your students’ names,” said Dukes.

According to Watkins, AI that provides feedback to students like OpenAI can prioritize white standardized vernacular in English, leaving out students who may speak and write from a different cultural framework. Students might also “have neurodiversity that requires a different level of nuance to be brought to the assessment process,” Watkins continued.

Even with an agreed upon set of policies and tools, change is inevitable. According to Dukes, the real challenge is that in a few years, once understanding of AI technologies is better, “then we might have a whole new generation of AI capabilities, and AI powered tools.”

Teachers are Still Hesitant About Using AI

For Marcus Luther, a high school English teacher in Oregon, AI’s implementation in the classroom and in K-12 teaching has moved too quickly. He doesn’t use AI in his lesson planning or classroom, and his current curriculum standards don’t require him to teach his students about AI use. He doesn’t feel confident enough with the ever-growing technology of generative AI to use it outside of curriculum standards in a thoughtful, ethical, and academically minded way.

He said he has had one professional development session to address AI tools for educators, but the approaches he’s seen haven’t made him feel supported in implementing AI in the classroom because of broader implications.

What he’s looking for is to deepen the process of learning and isn’t sure the tools he’s seen accomplish that, but may favor a “shortcut towards efficiency.”