A bipartisan group of U.S. senators has formally requested that the Government Accountability Office (GAO) investigate how artificial intelligence is affecting K-12 education, focusing on student achievement, teacher readiness, and special education outcomes. The move signals growing concern among lawmakers that AI adoption in schools is outpacing our understanding of whether these tools actually improve long-term learning.

What Does the Research Actually Show About AI Tutoring?

The senators' letter, led by Ranking Member Lisa Blunt Rochester (D-Delaware) and Chair Tommy Tuberville (R-Alabama) of the Senate Committee on Health, Education, Labor, and Pensions' Subcommittee on Education and the American Family, highlights a troubling gap in what we know about AI's educational value. Students are already using a wide range of AI tools in classrooms, including generative AI chatbots, tutoring assistants, writing helpers, and feedback systems that provide hints and explanations on assignments.

The promised benefit sounds compelling: AI can deliver personalized, real-time feedback as students work through math problems or essays. However, early research reveals a more complicated picture. While students may see short-term efficiency gains when using AI tutoring tools, it remains unclear whether these gains translate into deeper understanding or stick with students over time. The senators explained the concern plainly: a student using an AI math tutor might complete homework faster and feel more confident in the moment, but the reduced reasoning required by relying on AI hints can lead to weaker recall of the material later.

Why Are Teachers Unprepared for AI in the Classroom?

One of the most pressing issues the investigation will examine is teacher preparedness. A survey of 303 teachers across the United States and India revealed a stark reality: only 3 percent of teachers reported regularly integrating AI into their practice, while 37 percent described themselves as just beginning to explore the technology. Meanwhile, 24 percent reported creating specific AI activities, 17 percent had received training in AI, and 3 percent reported not using AI at all.

This fragmented adoption matters because the senators emphasized that AI tools are most effective when they augment rather than replace human instruction. Teachers need to understand not just how to use AI, but also how to supervise students using it safely and how to recognize when AI is appropriate for a learning goal and when it is not. Yet many educators feel unprepared for this role. A Pew Research Center study found that one quarter of surveyed U.S. teachers believe AI tools are doing more harm than good in K-12 education.

Steps to Ensure AI Supports Rather Than Undermines Teacher Effectiveness

• AI Literacy Training: Teachers need foundational knowledge about how AI works, what it can and cannot do, and how to evaluate AI tools before adopting them in the classroom.
• Supervision and Safeguards: Educators must learn how to monitor student use of AI tools, set appropriate boundaries, and ensure students are not bypassing critical thinking by over-relying on AI assistance.
• Ongoing Professional Development: As AI tools evolve, teachers need continuous support and updated training to select tools suited to specific learning objectives and to understand emerging risks like data privacy concerns.

The senators noted that while some teachers express enthusiasm about AI's potential to reduce workload during a time of widespread educator burnout, many others feel fear and apprehension about the technology. This mixed sentiment underscores why teacher preparation is central to whether AI can be used effectively to support student learning.

The investigation will also examine a particularly sensitive area: the use of AI in special education. AI-driven tools show promise for students with disabilities, including conversational chatbots for students with autism spectrum disorder, AI-annotated text for students with dyslexia, and speech synthesizers for students with speech and language impairments. However, these tools raise significant concerns about student data privacy and whether teachers are adequately trained to use them safely. There is also growing concern about the use of AI in drafting Individualized Education Programs (IEPs), which are legally binding documents that outline specialized support for students with disabilities.

What Questions Will the GAO Investigation Answer?

The senators have asked the GAO to investigate several critical questions, including what researchers have learned about AI's effects on students' critical thinking skills. This reflects a core worry: that while AI might help students complete assignments faster, it could undermine the deeper cognitive work that builds lasting knowledge and problem-solving ability.

The investigation comes at a moment when AI adoption in schools is accelerating without sufficient evidence of its long-term impact. The senators acknowledged that AI tools in education have promised benefits for both educators and students, but they stressed that early observations raise important concerns. As AI adoption in schools grows, understanding how these tools are being used and what effects they actually have is critical for making sound policy decisions in the future.

The formal GAO investigation signals that Congress is taking seriously the need to move beyond hype and anecdote to rigorous evidence about whether AI is truly improving K-12 education or simply changing how students and teachers work without delivering meaningful learning gains.