AI In Education: What’s Really Happening In US Schools In 2025

Although we at Third Space Learning are deep into the use of artificial intelligence or AI in education as we develop the capabilities of Skye, our AI tutor, we know some of you are still understandably cautious about AI integration in your schools and workflows.

In this article, we provide an overview of how AI is being used in schools around the US. The idea is to give you the encouragement to start your own experimentation or to suggest some additional pragmatic and teacher-tested ideas to deepen and extend AI’s role in your professional life.

The ideas and examples we use are drawn from our own usage as experienced educators and from the teachers and school administrators we are talking to each week, who are moving quickly from “Should we use AI?” to “How can we use AI effectively?“

AI in education

In US education, AI now automates administrative tasks and processes, enhances lesson plans, and personalizes tutoring. Done well, it can save teachers hours every week, boost student engagement, and meet FERPA-mandated safeguarding rules by keeping a human “in the loop.”

The current state of AI in education

In our 2025 standardized testing survey, we asked 6th grade teachers and elementary school administrators about their usage of AI as part of their math test preparation. Only 23.3% of the participants in the survey had used any form of AI (mostly ChatGPT or Teachmate).

However, a recent 2025 Twinkl survey of 6,500 teachers found 60% are using AI technologies for work purposes.

Back in 2023, research found that 42% of elementary and high school teachers were already using generative artificial intelligence (GenAI) in their educational role.

Of course, the truth is that we are all on our own varied paths toward the adoption of artificial intelligence in education.

Initial reluctance around AI in education

The adoption of AI in education has grown steadily, though not uniformly. 60% of teachers might be using it, but 76% report not receiving any training. Many AI technologies already automate routine tasks and administrative tasks, but a lack of training and confidence then becomes a common theme in teachers’ reluctance to have a go.

It appears there are other barriers to adoption, such as cost – many teachers have to pay, for example, for ChatGPT themselves – or just general skepticism on whether the answers are going to be correct or whether it’ll be just as quick and effective to use their usual learning resources.

We’ve already done a deep dive into how teachers are using AI in math, but more widely, this is what we see around artificial intelligence in education.

What is AI in education?

AI in education refers to the use of artificial intelligence technology to enhance the learning experience and streamline educational processes. This includes personalized learning through adaptive platforms, automated grading and scheduling, and interactive learning tools. AI offers significant potential for improving education, but also raises concerns about data privacy, equity, and the role of teachers.

AI adoption across the education sector

As you’d expect, some educational institutions have embraced AI systems enthusiastically while others approach with greater caution.

We are focusing this article on the use of AI in elementary and high schools, but we have much to learn, as always, from colleagues in early childhood and postsecondary education. To generalize broadly, these are some of the patterns we’ve observed within different education settings:

• Early childhood and preschool educators say the biggest gain from AI in their education setting is time; quicker evidence gathering means more time for child-led play. Creative ideas include using voice-prompted story generators, AI image cards for phonics, and automated observation notes inside pre-K tracking apps.
• Elementary schools are focusing on the basics: AI tools that help with planning, grading, and admin tasks, and lesson preparation. Teachers report spending more time connecting with struggling readers and less time formatting worksheets.
• High schools are getting more adventurous with subject departments experimenting with teacher- and student-facing AI. Science departments are using AI to help students design experiments and predict outcomes, transforming how they teach the scientific method.
• Community colleges and vocational schools are often where the most innovative practices emerge. Instructors describe using AI to simulate workplace scenarios, allowing healthcare students to practice difficult conversations before facing real situations.
• Universities are leading the way in artificial intelligence in education, especially in administration and accessibility. AI now adds transcripts, summaries, and quizzes to lecture recordings automatically. STEM labs use it to flag anomalous results, and disability services teams rely on AI note-takers to produce accessible materials quickly. It’s all about scale, helping multiple students without adding to staff workload.

The top 5 LLMs for education AI in schools

Before we look at specific educational AI tools, it’s worth understanding which of the main LLMs are available to teachers and what their benefits and use cases might be. We’ve not included DeepSeek on this list despite it having some fantastic reasoning capabilities, as there are still some data and privacy concerns here.

ChatGPT (OpenAI)

The tool that started the generative AI revolution in the education sector remains popular with teachers. Schools report that the free version is useful for basic tasks, while some departments invest in Plus subscriptions for more advanced capabilities.

Best for:

• Creative writing prompts;
• Explanation of concepts;
• Generating discussion questions;
• Quick resource creation.

Teachers particularly value ChatGPT’s ability to adapt its language to different age levels, making complex concepts accessible to younger students.

Want to know more? We share an array of effective timesaving prompts and ideas for using ChatGPT for math teaching here.

Limitations:

• Without careful prompting, responses can sometimes be too generic;
• Teachers report occasionally needing to ask for multiple iterations to get exactly what they need;
• For anything more than superficial exploration, it’s generally only worth using the paid-for version.

Microsoft Copilot (formerly Bing Chat)

Available through Microsoft 365, many schools are already discovering Copilot as part of their existing licenses. Integration with Office applications makes it particularly valuable for schools invested in the Microsoft ecosystem.

Best for:

• Document drafting and editing;
• PowerPoint creation;
• Data analysis in Excel;
• Email management.

High school teachers report that Copilot’s integration with existing workflows makes it particularly time-efficient.

Limitations:
Full functionality may require both Microsoft 365 and additional Copilot licenses, which can be a budget consideration, but it may already be part of your Microsoft suite of tools.

Google Gemini (formerly Bard)

Schools using Google Workspace are increasingly exploring Gemini, particularly for its integration with Google’s educational tools.

Best for:

• Collaborative document creation;
• Research summaries;
• Translation services;
• Image analysis.

Elementary school teachers report that Gemini’s ability to provide quick, simple explanations makes it valuable for supporting classroom teaching.

Limitations:

• The advanced features require a subscription;
• Some teachers report a steeper learning curve compared to other tools.

NotebookLM (Google)

This newer tool is gaining popularity, particularly among high schools, for its ability to work with multiple documents and research materials.

Best for:

• Research projects;
• Literature analysis;
• Helping students organize complex information.

Teachers value its ability to keep track of sources, making it especially useful for teaching proper research methods.

Limitations:

• Currently more specialized, with a focus on research and analysis rather than general-purpose applications.
• Early-adopter schools praise this aspect but still treat it as beta software.

Specialized educational AI technologies

Beyond these general tools, purpose-built educational AI is making a significant impact:

Third Space Learning’s Skye is a conversational AI math tutor for elementary and high school students. It adapts to individual student needs, using a voice-based interface that engages students in mathematical dialogue.

Century Tech offers personalized learning pathways across multiple subjects based on AI analysis of student performance to improve educational outcomes.

TeachMate AI focuses specifically on reducing teacher workload through lesson planning and resource creation.

Schools report that these specialized tools often integrate more seamlessly with educational needs, though general-purpose AI can offer greater flexibility across different tasks.

What about you?

Take 15 minutes to jot down your three biggest time-drains. Could AI help with any of them? Many teachers start with simple tools like navigation apps for organizing information before moving to more advanced AI applications. Focus on just one practical problem you’d love to solve.

AI in education examples

AI in education is used in various ways to enhance the learning experience and support educators. It can automate tasks, personalize learning, provide feedback, and even create virtual learning environments.  These are the most common examples of AI in education we see in US schools today.

Administrative systems and MIS integration

Perhaps the least glamorous but most immediately beneficial application of AI in education is in administrative systems. School leaders and support staff report significant time savings through:

• Intelligent dashboards that surface anomalies and patterns in attendance, behavior, and achievement over large volumes of data;
• Smart scheduling systems that optimize timetabling and resource allocation;
• MIS integrations that streamline reporting and reduce duplicate data entry;
• AI-assisted financial planning tools that help forecast budgets and identify potential savings

Practical examples of AI in school administration

• Special education coordinators are using AI to draft initial IEP suggestions based on assessment data.
• Schools are implementing AI tools that flag attendance patterns, allowing earlier intervention.
• Schools are using AI to help draft consistent yet personalized communications to parents.
• Arbor MIS has integrated AI analytics that automatically flag attendance patterns and suggest intervention strategies for persistent absenteeism.

Try this next: Ask your MIS provider if they have AI features you’re not using. Many systems now include predictive analytics and automated reporting tools that you might already be paying for.

Curriculum and lesson planning

The way most of us first start experimenting with AI in education tends to be in planning a lesson or creating a set of questions or a resource (see below) to enhance teaching. Teachers talk about:

• Generating differentiated lesson plans tailored to specific student needs;
• Suggesting creative activities and approaches for teaching challenging concepts;
• Curating relevant, curriculum-aligned resources from across the web;
• Adapting existing materials for different abilities and learning styles.

Practical examples of AI for curriculum and lesson planning

• Grades 3-6 teachers are using AI to generate varied math problems at multiple challenge levels.
• High school English departments have created banks of AI-generated questioning stems aligned to Bloom’s taxonomy. 
• Foreign language teachers are using AI to help create sample, levelled curriculum materials to improve language learning.
• AI can also auto-generate variation tables (e.g. 480 ÷ 10 → ÷ 0·1) so students spot the underlying structure.
• TeachMate AI offers personalized planning that adapts to specific class needs and curriculum objectives.

What’s interesting is how teachers are using AI not to replace their planning but to enhance it. The technology provides a starting point, which teachers then adapt with their knowledge of:

• What each learner needs;
• Classroom dynamics;
• Curriculum context.

Try this next: Paste this prompt into your LLM of choice: 

Generate 10 fluency, 10 reasoning and 10 problem-solving questions on equivalent fractions. Return in a simple table.

Review and tweak. Teach one set next week; log student responses.

Resource creation

Creating resources used to be a major time sink. Now, provided you give the AI enough context, it can halve the time you spend on this aspect of the job. Many of the educational AI tools can:

• Generate worksheets, quizzes, and assessment materials aligned to specific curriculum objectives;
• Create visual aids, diagrams, and graphic organisers;
• Develop scaffolded writing frames and prompts for different ability levels;
• Produce translations and simplified versions of texts for ELL students or for language learning.

Practical examples of AI for resource creation

• Geography teachers can create tailored worksheets on climate change in seconds.
• Elementary teachers are generating personalised reading comprehension questions for guided reading sessions.
• Drama teachers use AI to help create character profiles and scenario prompts.
• Math leads use AI-written GeoGebra scripts to spin up interactive fraction bars or dynamic graphs in seconds.
• Canva for Education now includes AI design assistants that help create visually engaging worksheets and presentations.

Teachers aren’t using AI to create perfect resources – they’re using it to create better first drafts that they can quickly customise and adapt into new content. The quality improves because they can focus on the teaching rather than the formatting.

Try this next: Choose one upcoming lesson and ask an AI tool to create a simple resource for it – maybe exit tickets, a starter activity, or a knowledge organizer. See how much time it saves you.

Personalized learning and tutoring

This is where we’re seeing some of the most exciting developments, including our own conversational AI tutor, Skye.

Teachers initially expressed skepticism about AI tutoring:

• Could AI technology really adapt to individual student needs?
• Could it spot misconceptions?
• Could it build the rapport necessary for effective learning?

Schools have discovered that while AI can’t replace a human teacher, it can provide personalized support at a scale that would otherwise be impossible with only human intelligence.

Special education co-ordinators tell us that AI tutoring allows all students needing additional support to receive one-to-one attention simultaneously. Teaching assistants can then focus on the students with the most complex needs.

The power isn’t just in the numbers, though. It’s in the adaptive nature of these systems that:

• Patiently work through the same concept multiple times without frustration;
• Interact with students, adjusting the level of challenge in real-time based on student responses;
• Provide immediate, non-judgmental feedback.

Teachers report that some students who are reluctant to participate in class or ask for help are more willing to engage with AI tutors, talking to it, taking risks and making mistakes – essential elements of effective learning.

Try this next: 

Think about the individual support each child in your class needs. One-to-one tutoring on a large scale might be unaffordable, but AI tutoring can provide personalized lessons to each child. 

What if you could ask an AI tutoring tool to run pre-teaching mini-sessions on next week’s class topics so students arrive primed for whole-class input?

Assessment and feedback

AI is significantly reducing teacher workload around school assessment. 

• Automating the marking of objective questions and tests;
• Providing initial feedback on written work that teachers can then refine;
• Identifying patterns in student errors to inform and enhance teaching;
• Tracking progress against curriculum objectives over time.

Teachers tell us that AI can handle the first stage of marking – spotting technical errors, checking facts, and suggesting improvements. This frees them to focus on the personal comments that show they really know the student.

Practical examples of AI for assessment and feedback 

• English teachers are using AI to provide initial feedback on writing, which they then refine.
• Science departments are using AI to mark multiple-choice quizzes and identify common misconceptions.
• Foreign language teachers are using AI to check translations and provide formative assessment feedback.
• Math departments deploy AI ‘step-analysis’ marking that pinpoints the line where an algebra solution goes off-track.
• AI builds error-analysis worksheets where students match common math misconceptions to corrected working.
• No More Marking’s Comparative Judgement uses AI to help standardize writing assessment.

Try this next: For your next set of student work, try having ChatGPT, Microsoft Copilot, or Google Gemini do an initial assessment before you review it. See if it catches things you might miss or offers useful feedback suggestions.

Professional development

School leaders are increasingly incorporating AI tools into their professional development strategies, recognising that personalized learning is as valuable for teachers as it is for students.

• Personalised CPD recommendations based on teaching goals and interests, for example AI for math;
• Virtual coaching that provides feedback on lesson delivery;
• Communities of practice enhanced by AI-curated research and resources;
• Simulation tools that allow practice of challenging classroom scenarios.

Practical examples of AI for professional development

• Newly qualified teachers and early career teachers are using AI to help reflect on lessons and suggest improvements.
• Schools are creating AI coaching assistants that help teachers implement strategies from CPD sessions.
• Schools are using AI to personalize professional development pathways based on teacher needs and interests.
• IRIS Connect has added AI features that analyse lesson recordings and suggest areas for development.

What’s particularly powerful is how AI can offer private, non-judgmental support. Teachers can ask for advice on challenging areas of practice without feeling embarrassed or exposed.

Try this next: Think about an area of teaching you’d like to develop. Ask an AI tool to create a personalized learning plan with practical strategies and resources.

Compliance and policy drafting

School leaders tell us that keeping policies up-to-date with ever-changing regulations is a significant administrative burden. AI is helping lighten this load:

• Principals are using AI to review existing policies against new guidance, highlighting areas needing updates;
• School business managers use AI to draft initial versions of compliance documents;
• School safety co-ordinators report using AI to create age-appropriate privacyand safety resources for different year groups;
• Special education co-ordinators are using AI to help keep IEPs aligned with changing statutory requirements.

One of the most valuable applications is policy gap analysis – using AI to compare existing school documentation against new frameworks like state education department standards or updates to Title IX and FERPA regulations.

Schools emphasise that human oversight remains essential – content produced by AI always needs careful review by knowledgeable staff.

Final policies must be signed off by a designated school official or school board member – AI drafts serve only as a starting point, not the finished document. However, having AI handle the initial heavy lifting can reduce policy development time by up to 70%.

Try this next: Run the school’s calculation policy through an AI readability check; flag sections above grade-appropriate reading levels.

Top 5 benefits of AI in education

Time-saving and workload reduction through automation

If there’s one benefit that resonates with every educator we talk to, it’s time. AI is giving teachers back hours in their week.

Teachers report finishing work earlier, sleeping better, and having more energy in the classroom. This isn’t just about teacher time and teacher wellbeing (though that’s crucial); it’s about educational quality and how it impacts students’ lives. Teachers who aren’t stretched thin have more creativity and enthusiasm to bring to their classrooms.

Personalization at scale

In every classroom, teachers are trying to provide tailored support that meets diverse needs with limited resources. AI is helping bridge this gap.

With classes of 25 to 35 students—often with a range of learning styles, IEPs, and English language levels—AI-powered tools are supporting differentiation more effectively than ever. Educators report using adaptive software, personalized practice tools, and AI tutors to tailor instruction, challenge high-achieving students, and provide scaffolding for those who need more support.

Accessibility and inclusion

For students with additional needs, AI tools are removing barriers to learning:

• Text-to-speech tools help dyslexic students access complex texts;
• Speech recognition enables students with physical disabilities to demonstrate their knowledge;
• Real-time translation supporting ELL students.

Special needs co-ordinators report that these tools are helping students with diverse needs participate more fully in mainstream education and demonstrate their true capabilities.

Enhanced learning experiences and engagement in learning

Beyond efficiency, AI is enabling new kinds of learning.

Science departments are using AI simulations to let students experiment with scenarios that would be too dangerous, expensive, or time-consuming to replicate in school. Student engagement in these virtual experiments is reportedly exceptional.

Data-driven insights

The analytical power of AI is helping schools spot patterns and intervene earlier.

Schools are discovering correlations between seemingly unrelated factors, like attendance patterns in one year and academic performance several terms later. This allows for more targeted, timely math interventions.

Challenges and concerns around AI in the education sector

While we’re excited about AI’s potential, we’re not wearing rose-tinted glasses. In our conversations with schools, several challenges consistently emerge:

Data privacy and protection

School leaders express legitimate concerns about student data privacy when implementing AI tools. Schools need clear guidance on:

• Which AI tools comply with FERPA (Family Educational Rights and Privacy Act);
• What types of student data can safely be processed;
• How to obtain appropriate permissions from families and guardians.

To address this, some schools are developing a whitelist of approved AI tools that have been vetted by their data privacy coordinator or IT director.

Compliance and governance

Beyond data privacy, schools are working to ensure that AI use aligns with broader educational policy, ethical practices, and student protection requirements. Key actions include:

• Creating clear acceptable use policies for AI in educational settings;
• Requiring human review of AI-generated content before it’s used with students;
• Ensuring district leadership teams and school boards understand enough about AI to provide appropriate oversight;
• Regularly auditing AI tools to ensure they continue to meet academic and ethical standards.

Districts with strong governance frameworks tell us they feel more confident experimenting with AI because they’ve established clear guardrails.

Keep an eye on federal and state-level guidance around AI use in schools — frameworks from the U.S. Department of Education and state education agencies are evolving.

Potential for bias and inaccuracy

AI models can reflect biases found in their training data. History teachers report that AI-generated suggestions may oversimplify or sanitize controversial topics, or present content with a skewed Western perspective.

Whatever your subject area, it’s critical to review AI-generated content with a critical lens — and to teach students to do the same.

The learning curve for educators

Many teachers want to use AI tools, but finding time to learn new tech on top of existing demands is a challenge. Schools seeing success often:

• Start small — with quick wins that show value immediately;
• Designate “AI champions” or instructional tech leads to support other staff;
• Build in time for trial-and-error use.

The learning curve for students

Just because students are digital natives doesn’t mean they know how to use AI well.

Middle and high school teachers report that some students initially used AI tools simply to “get the answers.” Schools have had to explicitly teach students to use AI for learning, not just shortcuts.

To build AI literacy, some schools are now teaching students how to:

• Write effective prompts;
• Evaluate AI-generated outputs critically;
• Use AI as a learning partner, not a replacement for their own thinking.

Maintaining the teacher-student relationship

School leaders voice concern that too much reliance on technology could erode the human connection at the heart of teaching.

But the most successful schools see AI as a support — not a replacement. When AI handles repetitive tasks (e.g., generating examples, summarising material), teachers have more time for meaningful interactions that build trust and curiosity.

Ethical considerations

The use of AI in education also raises ethical questions about the potential for manipulation, misinformation, and the impact on student autonomy.

Best practices for implementation

Through conversations with hundreds of schools at different stages of their AI journey, some clear patterns of success have emerged:

Strategic planning

Schools having the most success aren’t adopting AI indiscriminately. They’re starting with clear educational goals.

Effective schools identify specific challenges they want to address, such as how to reduce teacher workload, support for struggling readers, and provide challenges for the most able. They then research AI tools specifically for those purposes.

Teacher training and support

Schools report that effective AI training needs to be hands-on, not theoretical. Teachers need experience with the tools, ideally solving real problems from their practice.

Effective approaches include:

• Lunch-and-learn sessions where teachers share practical applications;
• Digital champions who provide peer support;
• Protected time for experimentation and reflection.

Starting small and scaling

Every successful school implementation we’ve seen started with a focused pilot.

Schools typically begin with just one team or department using AI for a specific purpose. As benefits become clear, other departments naturally become interested in adoption.

Measuring impact

District leaders emphasize the importance of evaluating AI investments. Schools are developing various approaches:

• Comparing teacher workload before and after implementation;
• Tracking student engagement and progress;
• Gathering qualitative feedback from all stakeholders;
• Any education artificial intelligence providers will be able to provide evidence of impact.

Involving all stakeholders

The most successful implementations involve the whole school community.

Schools report that transparency with parents about AI tools builds trust. Evening workshops demonstrating the technology in action can generate excitement rather than concern.

Try this next: Form a small working group of interested staff to explore one specific AI application. Give them the time and permission to experiment, fail, and learn together.

The future of AI in education: What’s on the horizon

Several trends seem likely to shape AI in education, not just in teaching and learning, but in other areas of school operation too:

Increasingly sophisticated tutoring systems

Our AI tutor, Skye, demonstrates how these intelligent tutoring systems are becoming more responsive to student thinking and developing increasingly intelligent behaviors. We expect tutoring services will continue to evolve, with more natural conversations, better identification of misconceptions, and smoother adaptation to individual learning patterns.

Skye is improving every day in how it responds to students and helps them to make accelerated progress.

Greater integration of AI in education

Rather than separate AI tools for different purposes, we’re likely to see more integrated systems that work within existing platforms. Teachers might plan lessons in their usual system with AI suggestions automatically appearing, or mark work with AI assistance built into familiar assessment platforms.

Enhanced collaboration between humans and artificial intelligence

The future isn’t AI replacing teachers; it’s AI and teachers working together in increasingly sophisticated ways. As one head of math put it: “The AI handles the repetitive tasks while I focus on the aspects of teaching that truly require human judgment and connection.”

Taking the first step toward integrating AI in your school or classroom

If you’ve read this far, you’re probably either already experimenting with AI in your school or seriously considering it. Here’s our advice, based on hundreds of conversations with educators at all stages of this journey:

• Start with a specific problem you want to solve. The most successful implementations address real needs, not abstract notions of “innovation.”
• Begin small and build confidence. Choose one application in one area of your school with motivated staff.
• Share successes and challenges openly. Create spaces for teachers to discuss what’s working and what isn’t.
• Keep students and learning at the center. The technology should serve your educational goals, not the other way around.
• Be patient with yourself and others. This is new territory for everyone. Mistakes and false starts are part of the learning process.
  Most importantly, take that first step. Schools that started exploring AI a year ago are now reaping significant benefits.

AI in education FAQs