Shaping a Better, More Equitable Future Using Artificial Intelligence
The ethical and equitable use of AI is integral to AIR’s corporate responsibility and organizational culture. We develop and implement new strategies to ensure fairness, transparency, and inclusivity in all our AI systems.
The transformative potential of artificial intelligence (AI) has enriched how we think about solving urgent social challenges.
AIR experts are constantly exploring new ways to apply AI and AI-enabled tools to solve real-world problems and improve lives. These tools can enhance our ability to automate tasks, inform decision making, and create innovative products and services that benefit society.
At the same time, we need more evidence about their efficacy and must actively work to better understand environmental impacts of AI use while mitigating risks such as bias and quality control. AIR is poised to fill this void by building robust evidence bases and implementing evidence-based practices for AI’s use.
At AIR, we believe that AI presents incredible promise for our overall mission to generate and use rigorous evidence that contributes to a better more equitable world.
Joseph P. Wilson
Managing Director
AIR understands the opportunity that AI provides and follows an evolving set of guardrails for its safe, ethical use in our work and in our organization. We are examining how the use of AI in research and technical assistance may improve how we generate and use evidence to improve lives. We are also considering novel applications of AI to improve our internal processes and benefit our clients and communities.
Notable Projects, Papers, and Initiatives
The following AI-related projects underscore AIR’s commitment to advancing the responsible, ethical use of AI in our work:
Exploring How AI Tools Can be Used to Improve Outcomes
Based on decades of research on human tutoring and AI in computer-based training systems, AIR, Per Scholas, and the University of Memphis are assessing the impact of intelligent tutoring systems, also called adaptive computer-based learning environments, on improving outcomes for diverse learners in Per Scholas’ best-in-class IT training program.
Operations Support for the Federal Health Insurance Marketplace
AIR uses AI tools in our operational support of the federal health insurance marketplace, which millions of consumers use to enroll in health insurance coverage. Large language models and other AI tools allow AIR to continuously streamline processes, increase accuracy, and implement efficiencies.
Effectiveness of AI-based Instructional Tools in Improving Undergraduate Biology Student Learning
This pilot program, funded by the National Science Foundation, is designed to expand knowledge about AI-based tutoring tools and their efficacy in undergraduate biology classrooms. AIR experts will conduct a meta-analysis of existing literature to understand AI’s effect on biology instruction and student learning outcomes.
Generating Evidence about AI Use in Research and Technical Assistance
Principles to Guide Artificial Intelligence in Education Research: Three Considerations
AIR researchers have summarized ongoing plans to develop a set of evolving principles for how to address key issues related to AI in education. This blog post highlights the importance of developing vetted uses of AI in education research methods as well as generating rigorous research evidence for AI-enabled teaching and learning applications as key to ensuring AI is a tool for positive change.
Advances and Challenges in Evaluating LLM-Based Applications
This paper describes innovative approaches to evaluating custom AI applications and their potential for creating barriers to faster progress in generative AI.
The Role of Artificial Intelligence in Workforce Development
This paper explores the role that advanced technology plays in creating flexible and affordable training programs necessary for building an inclusive, qualified workforce.
AI-Enabled Tools in Development
Teacher Support: Education leaders at AIR are building an AI-powered platform that will allow teachers to view metrics from observational videos recorded in their classroom. This platform will give teachers insight into their teaching practices on a regular basis, without the need for expensive in-person classroom observations.
EdTalk: EdTalk is a conversational tool designed to increase accessibility and user engagement for lengthy reports. AIR is responsible for overseeing all aspects of the project, including data preparation, model alignment via large language model prompting, model evaluation, and deployment.
