States Focused on Responsible Use of AI in Education during the 2025 Legislative Session

This blog is part of a series of pieces highlighting AI regulation trends across states during the 2025 legislative session. See CDT’s other blogs on public sector AI legislation and state administrative data privacy legislation.

Introduction

This year, the government and industry alike have given unprecedented attention to artificial intelligence (AI) adoption, implementation, and literacy in the education sector. In just one year, the White House issued an executive order on AI literacy and integration (Advancing Artificial Intelligence Education for American Youth); launched the Pledge to America’s Youth: Investing in AI Education that now includes commitments from over 200 tech companies and other organizations (including Google, OpenAI, Microsoft, NVIDIA, and Salesforce); the White House AI Task Force explicitly prioritized addressing deepfake non-consensual intimate imagery (NCII) including in schools; the U.S. Department of Education issued grantmaking guidance to improve educational outcomes using AI (Guidance on the Use of Federal Grant Funds to Improve Education Outcomes Using Artificial Intelligence (AI)); and the American Federation of Teachers launched the National Academy for AI Instruction in partnership with the United Federation of Teachers, Microsoft, OpenAI, and Anthropic.

At the same time, states across the country showcased the need to regulate the use of this technology in classrooms — during the 2025 legislative session 53 bills were proposed on the use of AI in education across 21 states, including 51 that apply specifically to the use of AI in K-12 schools. Out of these bills, 4 states across the political spectrum — Illinois, Louisiana, Nevada, and New Mexico — enacted legislation on AI in education. Additionally, 50 bills were proposed that focus more broadly on AI in the public sector, which can also apply to public education, and 79 bills and resolutions were introduced that address privacy and security protections for state administrative data, including education-related data.

Trends in 2025 State AI in Education Bills

Across the 53 bills that were proposed on AI in education in the 2025 legislative session, five trends emerge:

• Advancing AI literacy for students and professional development on AI for teachers;
• Requiring the creation of guidance or guidelines on the responsible use of AI in classrooms, including addressing issues like privacy, security, and transparency;
• Creating studies or tasks forces to assess the current state of AI in education and its impacts;
• Prohibiting specific AI uses in schools such as those related to student mental health support, student assessment, or replacing teachers; and,
• Addressing AI-generated NCII by updating cyberbullying policies and mandating appropriate response and prevention mechanisms.

Advancing AI literacy

Fifteen bills were proposed, one of which became law, focused on promoting AI literacy for students and professional development for teachers, making this the top priority of state legislators on the topic of AI in education. This recognized need for AI literacy aligns with nationally representative survey research from the Center for Democracy & Technology (CDT), which revealed that less than half of students and teachers have received some training or information about AI from someone at their school. Of those who did receive training, a majority found it helpful. 

One example of a bill proposed this session was New Jersey’s SB 3876, which would have: 

• Required school districts to incorporate AI literacy into existing curriculum for K-12 students (which should at a minimum include general concepts of AI, how to be an AI user, and best practices for responsible and ethical use of AI);
• Mandated that the state’s Commissioner of Education provide school districts with age appropriate sample learning activities;
• Required higher education institutions to offer certificate and degree programs in AI; and 
• Mandated that the state’s Secretary of Education (in consultation with other offices) develop an AI model curriculum for use by four year public higher education institutions and county colleges for the certificate and degree programs. 

Another example is Tennessee’s SB 677, which would have required the state Department of Education to provide a professional development course on the use of AI in the classroom, and mandated that teachers take such a course within two years of becoming a certified teacher.

Requiring the creation of guidance or guidelines on the use of AI in classrooms

The second largest category of bills introduced during the 2025 legislative session are the thirteen proposals, one of which was enacted, focused on requiring the creation of guidance or guidelines on use of AI in classrooms to address issues related to responsible use, including privacy, security, and transparency. New York’s AB 6972 would have created an AI working group to develop guidance and a model policy for school districts on responsible AI use and to assess the current state of AI in school districts and charter schools, including understanding the tools most commonly in use, their cost, how educators were involved in AI decisionmaking, and how AI is used in class content. Illinois’ HB 2503 would have established an Advisory Board to give guidance, oversight, and evaluation of education technology, including AI, in collaboration with the state education agency. The Board would have also been required to develop standards on topics like safety, privacy, and transparency for AI and guidance for school districts and educators on AI use and literacy.

Creating studies or task forces

Twelve bills were introduced, including one that became law, focused on creating studies or task forces to assess the current state of AI in education and its impacts. New Mexico enacted HB 2, which directs the legislative education study committee to create an education data governance and AI working group to study and recommend a structure for improved statewide education data governance and policies to address AI. Massachusetts’ SB 429 would have created a commission to study and make recommendations regarding AI in education, created guidelines for ethical AI use and suggest legislative proposals for student data protection, evaluated the current implementation and use of AI, and provided recommendations to the legislature on guardrails based on findings from a survey of the state’s school system, which includes evaluating aspects like the training teachers and students receive on AI and how the school system validates and tests AI systems on an ongoing basis.

Prohibiting specific AI uses

Eight bills were introduced, including one that was enacted, that would prohibit specific AI uses in schools such as those related to student mental health support, student assessment, or replacing teachers. For example, Nevada’s enacted AB 406 prohibits the use of AI in public schools for counselors, psychologists, or social workers (except for administrative efficiencies) and requires the state Department of Education to develop guidance for mental health professionals on how to responsibly use AI in providing services to students. The bill also provides that the guidance developed must give the Department of Education the authority to assess the effectiveness of tools used for these purposes. Related to this proposal, CDT research shows that only 11 percent of teachers say they have been trained on how to respond if they suspect a student’s AI use is detrimental to their well-being (e.g., hurting self-esteem, encouraging risky behavior). Another example is Texas’ HB 5282, which would have prohibited the use of AI to score responses to student assessments, unless a tool was trained on representative samples, demonstrates validity, and has been independently evaluated.

Addressing AI-generated NCII

States have responded to the growing threat of deepfake NCII by introducing five bills, one of which became law, that update cyberbullying policies and mandate appropriate response and prevention mechanisms. CDT’s research showed that the sharing of deepfakes and NCII remain prominent issues in K-12 schools. Twelve percent of students and 8 percent of teachers have heard of deepfake NCII in the last school year (2024-25) that depicts someone associated with their school. In response to this persistent threat, CDT created a model policy for K-12 schools to adopt to effectively address deepfake NCII. 

One example of a bill that addressed deepfake NCII this session is New Jersey’s AB 4736, which would have updated the state code to include deepfakes/deepfake NCII in the definition of harassment, intimidation, and bullying, and applied those provisions to teachers and staff. Similarly, Illinois enacted HB 3851, which updates the state’s school code to include the creation and distribution of AI-generated deepfake NCII in the definition of cyberbullying.

Strong Examples

Three proposed bills from the 2025 state legislative session stood out as strong examples for promoting responsible AI use in the classroom and placing appropriate guardrails on the technology:

• New York’s AB 6874A would have established an AI Literacy Digital Equity Competitive Grant Program to give grants to support AI literacy efforts; required award recipients to report information, such as the number of administrators or teachers who have been trained through the program, to the commissioner every year for four years; and required the state Education Department to develop a framework for evaluating the effectiveness of the grant program.
• Indiana’s HB 1296 would have required the state Education Department to create guidelines on school AI policies and a model policy, an inventory of AI platforms used in schools, a process and review by which teachers and school administrators can submit an AI platform for the inventory, conduct a survey of teachers and students regarding AI platforms, and submit a report of the survey results to the governor, legislative council, and members of the AI task force. The bill would have also required school corporations and charter schools to adopt, post, and communicate to students a school policy regarding AI and would have also prohibited school corporations, charter schools, and employees of a school corporation or charter school from penalizing a student based solely on evidence provided by an AI detector.
• Texas’ HB 1405 would have updated the definition of cyberbullying to include deepfakes/deepfake NCII, making it fall under legally required response and support mechanisms, including establishing investigative procedures for bullying incidents. The bill would have also required the district board of trustees to provide supportive services and counseling to victims of deepfakes or deepfake NCII and required schools and school districts to report the number of civil rights violations related to bullying incidents.

Conclusion

There is unprecedented federal attention to accelerating the deployment of AI in K-12 schools, and at the same time, there is growing evidence that this expansion is related to increased risks to students. Given this, and the likely continued investment into AI in education, it is critical that states create appropriate guardrails and provide guidance to teachers, students, and parents to ensure AI is being used in responsible, rights respecting ways. This is especially important given that some AI systems have failed to live up to their promises in the education context and have created new challenges or issues over the last few years.