Establishing a National Center for AI in Education

0

There are immense opportunities associated with artificial intelligence (AI), yet it is important to vet the tools, establish threat monitoring, and implement appropriate regulations to guide the integration of AI into an equitable education system. Generative AI in particular is already being used in education, through human resource talent acquisition, predictive systems, personalized learning systems to promote students’ learning, automated assessment systems to support teachers in evaluating what students know, and facial recognition systems to provide insights about learners’ behaviors, just to name a few. Continuous research of AI’s use by teachers and schools is important to ensure AI’s positive integration into education systems worldwide is crucial for improved outcomes for all. 

Congress should establish a National Center for AI in Education to build the capacity of education agencies to undertake evidence-based continuous improvement in AI in education. It will increase the body of rigorous research and proven solutions in AI use by teachers and students in education. Teachers will use testing and research to develop guidance for AI in education.

Challenge and Opportunity

It should not fall to one single person, group, industry, or country to decide what role AI’s deep learning should play in education—especially when that utility function will play a major role in creating new learning environments and more equitable opportunities for students. 

Teachers need appropriate professional development on using AI not only so they can implement AI tools in their teaching but also so they can impart those skills and knowledge to their students. Survey research from EdWeek Research Center affirms that teachers, principals, and district leaders view the importance of teaching AI. Most disturbing is the lack of support and guidance around AI that teachers are receiving: 87% of teachers reported receiving zero hours of professional development related to incorporating AI into their work. 

A National Center for AI in Education would transform the current model of how education technology is developed and monitored from a “supply creates the demand system” to a “demand creates the supply” system. Often, education technology resources are developed in isolation from the actual end users, meaning the teachers and students, and this exacerbates inequity. The Center will help to bridge the gap between tech innovators and the classroom, driving innovation and ensuring AI aligns with educational goals.

The collection and use of data in education settings has expanded dramatically in recent decades, thanks to advancements in student information systems, statistical software, and analytic methods, as well as policy frameworks that incentivize evidence generation and use in decision-making. However, this growing body of research all too frequently ignores the effective use of AI in education. The challenges, assets, and context of AI in education vary greatly within states and across the nation. As such, evidence that is generated in real time within school settings should begin to uncover the needs of education related to AI. 

Educators need research, regulation, and policies that are understood in the context of educational settings to effectively inform practice and policy. Students’ preparedness for and transition into college or the workforce is of particular concern, given spatial inequities in the distribution of workforce and higher-education opportunities and the dual imperatives of strengthening student outcomes while ensuring future community vitality. The teaching and use of AI all play into this endeavor.

An analog for this proposal is the National Center for Rural Education Research Networks (NCRERN), an Institute of Education Sciences research and development center that has demonstrated the potential of research networks for generating rigorous, causal evidence in rural settings through multi-site randomized controlled trials. NCRERN’s work leading over 60 rural districts through continuous improvement cycles to improve student postsecondary readiness and facilitate postsecondary transitions generated key insights about how to effectively conduct studies, generate evidence, influence district practice, and improve student outcomes. NCRERN research is used to inform best practices with teachers, counselors, and administrators in school districts, as well as inform and provide guidance for policymaking on state, local, and federal levels.

Another analog is Indiana’s AI-Powered Platform Pilot created by the Indiana Department of Education. The pilot launched during the 2023–2024 school year with 2,500 teachers from 112 schools in 36 school corporations across Indiana using approved AI platforms in their classrooms. More than 45,000 students are impacted by this pilot. A recent survey of teachers in the pilot indicated that 53% rated the overall impact of the AI platform on their students’ learning and their teaching practice as positive or very positive. 

In the pilot, a competitive grant opportunity funds the subscription fees and professional development support for student high dosage tutoring and reducing teacher workload using an AI platform. The vision for this opportunity is to focus on a cohort of teachers and students in the integration of an AI platform. It might be used to support a specific building, grade level, subject area, or student population. Schools are encouraged to focus on student needs in response to academic impact data. 

Table of Contents

Plan of Action

Congress should authorize the establishment of a National Center for AI in Education whose purpose is to research and develop guidance for Congress regarding policy and regulations for the use of AI in educational settings. 

Through a competitive grant process, a university should be chosen to house the Center. This Center should be established within three years of enactment by Congress. The winning institution will be selected and overseen by either the Institute of Education Sciences or another office within the Department of Education. The Department of Education and National Science Foundation will be jointly responsible for supporting professional development along with the Center awardee.

The Center should begin as a pilot with teachers selected from five participating states. These PK-12 teachers will be chosen via a selection process developed by the Center. Selected teachers will have expertise in AI technology and education as evidenced by effective classroom use and academic impact data. Additional criteria could include innovation mindset, willingness to collaborate with knowledge of AI technologies, innovative teaching methods, commitment to professional development, and a passion for improving student learning outcomes. Stakeholders such as students, parents, and policymakers should be involved in the selection process to ensure diverse perspectives are considered. 

The National Center for AI in Education’s duties should include but not be limited to:

  • Conducting research on the use of AI in education and its impact on student learning. This research would then inform guidance pertaining to development, use, policy, and regulation development. Potential research areas include:
    • Assessing learning: Process vs. Product. Plagiarism because of the Large Language Model’s (LLM) ability to produce text.
    • Use of generative AI as the first draft and beginning for creativity, thus raising the bar for student learning.
    • AI hallucinations.
  • Recommending AI technology for educational purposes, with a focus on outcomes. 
  • Establishing best practices for educators to address the pitfalls of privacy, replication, and bias.   
  • Providing resources and training for teachers about and use of AI in education.

Congress should authorize funding for the National Center for AI in Education. Funding should be provided by the federal government to support its research and operations. Plans should be made for a 3–5-year pilot grant as well as a continuation/expansion grant after the first 3–5-year funding cycle. Additional funding may be obtained through grants, donations, and partnerships with private organizations.

Reporting on progress to monitor and evaluate the Center’s pursuits. The National Center for AI in Education would submit an annual report to Congress detailing its research findings, advising and providing regulatory guidance, and impact on education. There will need to be a plan for the National Center for AI in Education to be subject to regular evaluation and oversight to ensure its compliance with legislation and regulations.

To begin this work of the National Center for AI in Education will:

  1. Research and develop courses of action for improvement of AI algorithms to mitigate bias and privacy issues: Regularly reassess AI algorithms used in samples from the Center’s pilot states and school districts and make all necessary adjustments to address those issues.
    1. Incorporate AI technology developers into the feedback loop by establishing partnerships and collaborations. Invite developers to participate in research projects, workshops, and conferences related to AI in education. Research and highlight promising practices in teaching responsible AI use for students:  Teaching about AI is as important, if not more important, as teaching with AI. Therefore, extensive curriculum research should be done for teaching students how to ethically and effective use AI to enhance their learning. Incorporate real-world application of AI into coursework so students are ready to use AI effectively and ethically in the next chapter of their postsecondary journey.
  2. Develop an AI diagnostic toolkit: This toolkit, which should be made publicly available for state agencies and district leaders, will analyze teacher efficacy, students’ grade level mastery, and students’ postsecondary readiness and success. 
  3. Provide professional development for teachers on effective and ethical AI use: Training should include responsible use of generative AI and AI for learning enhancement. 
  4. Monitor systems for bias and discrimination: Test tools to identify unintended bias to ensure that they do not perpetuate gender, racial, or social discrimination. Study and recommend best practices and policies. 
  5. Develop best practices for ensuring privacy: Ensure that student, family, and staff privacy are not compromised by the use of facial recognition or recommender systems. Protect students’ privacy, data security, and informed consent. Research and recommend policies and IT solutions to ensure privacy compliance. 
  6. Curate proven algorithms that protect student and staff autonomy: Predictive systems can limit a person’s ability to act on their own interest and values. The Center will identify and highlight algorithms that are proven to not jeopardize our students or teachers’ self-freedom.

In addition, the National Center for AI in Education will conduct five types of studies: 

  1. Descriptive quantitative studies exploring patterns and predictors of teachers’ and students’ use of AI. Diagnostic studies will draw on district administrative, publicly available, and student survey data. 
  2. Mixed methods case studies describing the context of teachers/schools participating in the Center and how stakeholders within these communities conceptualize students’ postsecondary readiness and success. One case study per pilot state will be used, drawing on survey, focus group, observational, and publicly available data. 
  3. Development evaluations of intervention materials developed by educators and content experts. AI sites/software will be evaluated through district prototyping and user feedback from students and staff. 
  4. Block cluster randomized field trials of at least two AI interventions. The Center will use school-level randomization, blocked on state and other relevant variables, to generate impact estimates on students’ postsecondary readiness and success. The Center will use the ingredients methods to additionally estimate cost-effectiveness estimates. 
  5. Mixed methods implementation studies of at least two AI interventions implemented in real-world conditions. The Center will use intervention artifacts (including notes from participating teachers) as well as surveys, focus groups, and observational data. 

Findings will be disseminated through briefs targeted at a policy and practitioner audience, academic publications, conference presentations, and convenings with district partners. 

A publicly available AI diagnostic toolkit will be developed for state agencies and district leaders to use to analyze teacher efficacy, students on grade level mastery, and students’ postsecondary readiness and success. This toolkit will also serve as a resource for legislators to keep up to date on AI in education. 

Professional development, ongoing coaching, and support to district staff will also be made available to expand capacity for data and evidence use. This multifaceted approach will allow the National Center for AI in Education to expand capacity in research related to AI use in education while having practical impacts on educator practice, district decision-making, and the national field of rural education research. 

Conclusion

The National Center for AI in Education would be valuable for United States education for several reasons. First, it could serve as a hub for research and development in the field, helping to advance our understanding of how AI can be effectively used in educational settings. Second, it could provide resources and support for educators looking to incorporate AI tools into their teaching practices. Third, it could help to inform future policies, as well as standards and best practices for the use of AI education, ensuring that students are receiving high-quality, ethically sound educational experiences. A National Center for AI in Education could help to drive innovation and improvement in the field, ultimately benefiting students and educators alike.

This idea is part of our AI Legislation Policy Sprint. To see all of the policy ideas spanning innovation, education, healthcare, and trust, safety, and privacy, head to our sprint landing page.

What is the initial duration of the proposed project?

Three to five years for the pilot, with plans developed for another three-to-five-year continuation expansion.

What is the estimated initial budget request?

$10 million. This figure parallels the funding allocated for the National Center for Rural Education Research Networks (NCRERN), a project of similar scope.

Why should a university house the Center?

Universities have the necessary capabilities to conduct research and to help create and carry out professional development programs. Additionally, this research could inform teacher preparation programs and the data disseminated across teacher preparation programs.

How would this new Center interact with the EdSafeAI alliance or similar coalitions?

The National Center for AI in Education would share research findings widely with all organizations. There could also be opportunities for collaboration.

Would the Center supplant the need for those other coalitions?

No. The Center at its core would be research-based and oriented at street level with teachers and students where the data is created.

link

Leave a Reply

Your email address will not be published. Required fields are marked *