How teachers are using AI
The First Avenue Elementary School in Newark, New Jersey is trying something new. The school is piloting Khanmigo, an AI-powered tutor and teacher support tool developed by Khan Academy.
Leticia Colon, an eighth-grade algebra teacher, is using AI to develop problem sets that feature hometown heroes and role-models, like Newark boxer Shakur Stevenson and his workout routines.
Cheryl Drakeford, a third-grade math and science teacher, uses Khanmigo to generate a first draft of rubrics and lesson hooks for assignments, which she then tailors with examples she knows her students will love. “Khanmigo gives me the blueprint, but I have to give the delivery,” she said.
There’s been plenty of attention on how AI could unlock new learning opportunities for students, help them cheat, or write papers, but what hasn’t been explored as much is AI’s impact on teachers, and if AI needs to be regulated in schools.
In this week’s newsletter, we explore how teachers are using AI, if the technology can be transformational for the profession, and the complex issues around artificial intelligence in the education system.
// AI: the new teaching assistant
Teachers at all grade-levels are using AI to stay ahead of the work and deliver better educational experiences for students.
- Lesson planning: New AI platforms provide tools for teachers to quickly generate lesson plans, tailor content to student interests, and generate first drafts that teachers can tweak.
- Grading essays and assignments: The AI tool Writable helps teachers grade writing assignments by delivering “rubric-aligned AI suggested comments and scores that drive more targeted revision.” AI can also be used to evaluate if students are using AI to cheat, though it’s not perfect, and there are questions about how AI is ethically and equitably grading student assignments.
- Reducing workload: The MagicSchool platform brands itself as an all-in-one AI-platform for teachers that can save them over ten hours a week through AI-generated writing, planning, and teacher training. For example, it will write the first draft of Individualized Education Programs (IEPs) for a teacher to finalize.
- Classroom management: An in-classroom AI voice-assistant like Merlyn Mind doesn't replace teachers, but it helps them have greater command of their classroom by summoning teaching materials and answering student questions.
- Creating “AI twins” of teachers: Two university professors in Lithuania have created AI chatbot versions of themselves that were trained on course material so students had access to 24/7 office hours. More and more teachers are relying on AI agents and AI tutors.
A 2023 study by The Walton Family Foundation found that ChatGPT is used by teachers more than students (at just over a year old, this study is already dated). Edtech companies see huge market potential: selling their latest tech to educators and school districts. Meanwhile, nonprofits like aiEDU are training teachers on how to apply the best of AI in all aspects of their work.
//
98% of teachers expressed the need for some level of education on the ethical usage of AI.
//
// Is AI a pipe-dream for educators?
Even outside of education, it seems like there is an AI solution for everything. But can these tools address the fundamental problems facing educators? According to a 2024 report by Rand on K-12 public school teacher wellbeing, 60% of teachers report being burned out, and 22% said they intended to leave their job at the end of the academic year. The top three reasons for teacher burnout:
- Managing student behavior.
- Low salaries (average annual base pay for a public school K-12 teachers is $70,000).
- The administrative work outside of actual teaching.
AI might be able to help streamline administrative work, but it seems unlikely that AI will solve the deeper, structural problems in the education profession. There is also the challenge that developing AI literacy becomes yet another responsibility for teachers who are already overworked (on top of it all, teachers need to ensure their AI models don’t hallucinate). Today, the average teacher uses over 140 different digital tools throughout the school year.
There’s also the equity question: as more AI tools that help teachers and offer personalized learning to students are paywalled, there will be greater disparities between the schools and school districts who can afford AI tools and those who cannot.
// The need to regulate AI use in education
The education sector is another domain where regulation is trying to catch up with fast-moving tech.
In the US, states are beginning to recognize the need to offer guidance to educators about how to use AI. For example, the California Department of Education doesn’t know which of its schools use AI or how much they pay for it. It also doesn’t track AI use by school districts, but that is beginning to change.
California is one of 23 states that has issued guidance on how to use AI in schools. The guidance makes recommendations on algorithmic biases, safe use, lesson planning, and personalized learning. But state guidance is not mandatory or enforceable; many schools and districts are left up to themselves to implement and monitor how AI tech is used.
At the federal level, the US Department of Education published a guide over the summer designed to aid the development of edtech tools that use AI. As more AI tools roll out, there will be greater pressure on state and federal departments of education to offer guidance and support to schools and educators.
Responsible AI development and regulation can borrow lessons from the efforts to make social media platforms safer and more private for users: from greater regulation at the state and federal level to greater individual agency and voice in shaping AI algorithms.
// Global perspective
Outside of the US, AI possesses similar opportunities and challenges. An analysis by the Bill and Melinda Gates Foundation identified three opportunities in education where AI solutions can be transformational:
- AI can expand access to high-quality learning resources that are tailored to varying needs, interests, and learning levels.
- AI can make teachers better teachers by giving them training, knowledge, and experience.
- AI can reduce the cost and time to develop high-quality educational content, as well as assessment and evaluation tools, in multiple local languages and contexts.
// Pencils down, chatbots up
AI tools are not panaceas for the complexity of issues facing today’s educators. For now, being a teacher at a time of rapid technological change requires navigating a variety of interconnected challenges: how to manage screens in classrooms (and varying parent perspectives on how screen-centric education should be); how to police AI usage so students aren’t outsourcing their critical thinking; and how to tap into the benefits of AI themselves. AI is here to stay, and teachers are on the front-lines of navigating it.
A report from 2023 found that 98% of teachers expressed the need for some level of education on the ethical usage of AI. After a period of every-school-for-themselves testing of AI, it will be important to verify that student data is protected and AI usage doesn’t exacerbate educational inequities.
As we have seen at Project Liberty, it is often harder to wrangle existing tech tools to protect data privacy than to incorporate those principles from the beginning with safety and privacy by design.
