Most teachers still use AI for basic tasks like lesson planning, but a growing movement is teaching educators to build more sophisticated tools that actually improve how they teach. About 50 teachers gathered in New York City on March 18, 2026, to learn how to develop what experts call "agentic" AI tools, autonomous software systems that can handle complex, multi-step tasks requiring reasoning to support instruction across subjects and grade levels. The shift reflects a broader maturation in how schools are adopting artificial intelligence. While 6 in 10 teachers now use AI in their practice, most still rely on it for surface-level work like creating lesson plans or managing calendars. The new training initiative aims to change that by helping educators understand how to direct AI tools using their own professional judgment and expertise, rather than simply accepting whatever the technology produces. What's Driving This Change in Teacher AI Training? The National Academy for AI Instruction, a five-year, $23 million partnership between the American Federation of Teachers and three major AI developers, Anthropic, Microsoft, and OpenAI, is at the center of this shift. The program aims to train 400,000 teachers on how to use AI technology in the classroom, with a focus on moving beyond basics. "We're in this race for teachers to get this knowledge of more meaningful use of AI," said Randi Weingarten, the AFT president. "This will become the most disruptive technology in our time. There is a real demand from educators to learn so that they are in the driver's seat for AI as opposed to the companies or districts or the technology itself". This training model is distinctive because teachers, with limited support from developers, train other teachers in how to use AI to improve instruction. The approach recognizes that educators understand their own classrooms better than any external trainer could. How Can Teachers Use "Agentic" AI Tools More Effectively? - Content Gap Detection: Teachers can use AI agents to stress-test lessons for content gaps and confusing wording, then refine their teaching approach over time based on what the AI identifies as problem areas. - Differentiated Instruction: Special education teachers can create and monitor individualized education programs and adjust lessons based on students' daily progress throughout the week, tailoring instruction to diverse learning needs. - Parent Communication: AI agents can draft personalized letters to families about field trips, academic challenges, or discipline issues while following district guidelines and incorporating research-backed resources and next steps. - Behavioral Problem-Solving: Teachers can use AI agents to brainstorm alternative approaches when a lesson or intervention isn't working, reducing the mental load of figuring out solutions on their own after school hours. - Culturally Responsive Teaching: Teachers can develop questioning techniques tailored to students from a wide variety of cultural backgrounds, helping them engage learners more effectively. The key advantage of agentic AI tools is that they allow teachers to narrow the scope of information the AI uses to respond to requests, according to Seth Reznik, a team lead for Microsoft Elevate, the tech giant's $4 billion training initiative for schools and nonprofits. This makes it less likely the AI will hallucinate, or make up information, and more likely to provide responses based on up-to-date research and trusted sources. Jing Liang Guan, a science teacher at the Brooklyn Science and Engineering Academy, explained the practical difference: "This means the difference between a tool that can create a lesson plan and one that can help him stress-test his lessons for content gaps and confusing wording, and help him hone his teaching approach over time". What Are Teachers Most Concerned About? Despite the promise of these tools, teachers are raising serious concerns about data privacy and student safety. Jennifer Watters, a 3rd grade teacher at PS 29 in Queens, recently switched from using OpenAI's ChatGPT to Anthropic's Claude specifically because of privacy concerns. "It's really important that teachers know that this information and these tools, if gotten into the wrong hands, can be very dangerous for our students, for our profession, and for our jobs," Watters said. Teachers also face a careful balancing act when using AI with sensitive student information. Yasheema Cook, who is developing AI agents to help her create individualized education programs for 12th graders in her self-contained special education classes, noted the challenge: providing enough context to the AI about student disabilities and needs to get meaningful help while also protecting student privacy. Additionally, some experts urge caution about offloading common teacher tasks to AI agents. While an AI agent that learns to write parent letters by studying examples of a teacher's actual letters can be efficient, it can paradoxically make teachers less likely to remember important details about their students. The broader concern about AI's impact on critical thinking persists. According to a Center for Democracy and Technology report, 70% of teachers worry that AI weakens critical thinking and research skills, and over half of students agree that using AI in class makes them feel less connected to their teachers. How Is AI Adoption Reshaping Education Overall? The scale of AI adoption in education is striking. According to a 2025 Microsoft report, 86% of education organizations now use generative AI, the highest adoption rate of any industry. The global AI education market reached $7.57 billion USD in 2025 and is projected to exceed $112 billion USD by 2034. The impact on student outcomes has been measurable. An AIPRM report of U.S. students found a 62% increase in test scores among those using AI-powered instruction systems, attributed to the technology's ability to identify and address knowledge gaps before they develop into larger challenges. AI-powered instruction platforms like Kyron Learning have shown up to 70% higher course completion rates compared to traditional approaches. At the same time, 69% of teachers said AI tools have improved their teaching methods, while 55% agreed that it has given them more time to interact directly with students. This suggests that when used thoughtfully, AI can actually free educators to focus on the human aspects of teaching that matter most. Rajen Sheth, Founder and CEO of AI-powered edtech company Kyron Learning, which raised $14.6 million USD in a Series A funding round, emphasized this point: "At the core, education remains human. The mentorship, connection, and inspiration that great educators provide cannot be replicated. The future is educators empowered by AI, not overshadowed by it". As 2026 unfolds, the challenge for schools will be ensuring that teachers receive the training and support they need to use these tools effectively and ethically. Randi Weingarten noted that "there's still a lot of fear in the absence of federal guardrails on privacy, on safety, on disinformation, on academic freedom". The National Academy for AI Instruction represents one significant effort to address that gap, but experts agree that ongoing training and guidance will be essential as AI technology continues to evolve.
