The U.S. is at a crossroads. Our current assessment systems most often deliver final scores, often gleaned through multiple choice, lagging woefully behind the complex, future-ready skills that matter most. This measurement gap prevents us from realizing the vision of assessment as a continuous, real-time guide for improvement. The good news? AI-powered measurement innovation makes this vision increasingly possible. Our critical challenge is ensuring the quality, explainability, and fairness of these new tools. Securing the future of learning and work depends on it. The Study Group works to elevate this national mission by advancing the best in AI, assessment, and data practice. 

Over the last 24 months, The Study Group has been leading a national initiative to transform education and workforce development through Artificial Intelligence (AI)-powered measurement and assessment innovation. This initiative—Advancing AI, Measurement and Assessment System Innovation—explores how AI might enhance learning, assessment, and educational efficacy by leveraging frontier technology, strengthening public-private partnerships, and ensuring that AI-driven tools provide meaningful, context-sensitive insights for learners, educators, workers, and families.

U.S. education and workforce systems are at a crossroads. While AI advancements offer unprecedented opportunities to personalize learning and improve assessment quality, existing systems remain outdated—overemphasizing easily measurable skills while neglecting durable, future-ready competencies. Federal investment in Public Infrastructures for AI in Education, including research and development, is crucial to ensuring that technology enhances learning, promotes opportunity, and meets evolving workforce demands. With broad momentum and a national shift toward AI use and integration, now is the time for strategic action.

The Study Group’s AI-driven R&D roadmap focuses on three core objectives:

• Strengthening Public-Private Partnerships – Aligning federal, philanthropic, and private-sector investment to drive high-impact AI and assessment innovation.
• Advancing AI in Learning & Assessment – Utilizing machine learning, generative AI, and advanced data analytics to create context-sensitive, personalized assessments that support all learners and workers.
• Securing Responsible AI Integration – Developing trustworthy, bias-free AI applications that protect student privacy and enhance instructional impact.

These benefits will enable: 

• Quality, Useful Assessments: AI has the potential to make assessments more engaging and useful for instructional decisions, providing reliable data to learners, educators, and families.
• Appropriate and Sufficient Supports for Learners: Calibrating easy-to-use tools and timely data to help educators rapidly support students.
• Balanced Measures: Encouraging state, school system, and campus leaders to value meaningful indicators to better guide school improvement, including school climate, student engagement, and skills for the future.

Innovating Assessment 

We aim to reimagine what we measure, how we measure, and for whom we measure while upgrading quality and explainability for AI-driven tools that provide timely, actionable feedback for students, educators, and families.

Innovating For Whom We Measure

In an era of multi-modal AI, the opportunity exists to center the needs of learners, educators, families, and the wider community in ways unthought of previously. In addition to addressing how the sciences of learning, development, measurement, and improvement are being leveraged to fuel innovation, our research surfaces lessons in Usefulness & Usability and Data-Driven Feedback & Decisions. Usefulness here refers to how well an assessment’s results inform real-world decisions, and usability encompasses the ease and clarity with which educators, learners, and other stakeholders can administer, interpret, and act on assessment results.

Innovating What We Measure 

With emerging scientific and technological breakthroughs, there is an opportunity to dramatically expand the field’s ability to measure complex yet vital competencies and contexts. Assessments signal priorities and focus thinking about what’s important, influencing what’s taught and learned. Our research elevates efforts to incorporate Broader, Holistic Definitions of Learning that view student growth not only as academic achievements, but also as social, emotional, and metacognitive development. It highlights the importance of real-world relevance, personal agency, and the interconnectedness of different subject areas. It also explores Next Generation Insights for Math & ELA, and how assessments are integrating conceptual understanding with procedural fluency, as well as framing problem-solving in authentic, real-world scenarios for greater engagement.

Innovating How We Measure

Measuring what matters most requires unlocking new advances in measurement, AI, design, science, and technology. These innovations support the design of tasks that capture rich, meaningful data without disrupting learning. Insights related to this theme touch upon the following innovations:

• Learner Variation & Person-Specific Assessment that incorporates each student’s unique profile for more accurate measures of progress and more finely-tuned feedback.
• Multi-Modal AI-Powered Approaches that combine various data inputs—such as text responses, speech, images, video, sensors, and even interactions—to form a richer picture of learner understanding and amplify opportunities for creative expression, collaboration, and real-time support.
• Competency-Focused Approaches (e.g., badges, portfolios, games, performance tasks) that emphasize mastery of specific skills or knowledge areas as demonstrations of proficiency, making progress more transparent and personalized.
• Participatory, Evidence-Centered AI Design that features collaboration between educators, students, and technologists to ensure assessment tools reflect real-world classroom needs and values.

Upgrading Quality & Explainability

As innovative technologies (including multi-modal AI) transform assessment, the field must remain attuned to the notion of quality and to acting in ways that grow the credibility and informed embrace of new assessments. Our research focuses on particular components of quality as defined by K-12 user-informed principles, including Scientific Soundness in the AI era and Transparency & Trust. This work touches upon scientific soundness, addressing the validity and appropriateness of analyses for particular purposes, as well as technical quality for the soundness of inferences about students. Research related to transparency and trust highlight considerations for ensuring that stakeholders understand how assessment data are collected, analyzed, and reported, and also how the upholding of privacy, fairness, and ethical safeguards can build trust.

Use Cases & AI Applications for Consideration

At the 2025 National Council on Measurement in Education Artificial Intelligence in Measurement and Education Conference (AIME-Con) this month, key applications worked through include: 

• Automated & Adaptive Assessments – AI-powered tools that adjust difficulty levels in real time and provide personalized feedback to students.
• Machine Learning in Large-Scale Assessments – Applying ML models to national assessments (e.g., NAEP) to improve scoring accuracy and insight generation.
• AI in Early Literacy & Text Complexity – Leveraging generative AI to tailor reading materials and ensure students receive appropriately challenging content.
• AI-Powered Writing Feedback – Automating structured feedback to enhance writing development and critical thinking.
• Dynamic Task Generation at Scale – Utilizing generative AI to create abundant, equivalent variations of assessment tasks, thereby enhancing test security, reducing cheating opportunities, and supporting extensive practice without repetition.
• Conversational Assessment Interfaces – Employing AI to conduct real-time, interactive dialogues with learners to probe understanding and fluency.
• Process-Based Assessment Analytics – Analyzing fine-grained observational data, such as keystrokes, eye-movements, revision patterns, time allocation, and problem-solving strategies, to reveal student thinking and metacognitive approaches beyond the final response.
• Operationalization of Complex Skills – Designing interactive and performance-based tasks and analyzing evidence (including group contributions) to measure critical durable skills such as collaboration, communication, and complex problem-solving that traditional tests often miss.
• Multimodal Performance Evaluation – AI analysis of student work that spans multiple formats, including text, images, audio, and video, to assess complex projects and real-world demonstrations of learning such debates or mock deliberations.
• Customization and Contextualization – Generating assessment scenarios embedded in realistic, learner relevant contexts based on student interests or background knowledge, promoting fairness, equity, and measuring the transfer of learning to real-world applications.
• Continuous Learning Progression Mapping – Continuously assessing student knowledge over time and across contexts (in and out of school) to generate detailed maps of learning, moving toward comprehensive learning profiles.
• Assessment of Human-AI Collaboration Skills – Evaluating students’ proficiency in working effectively with AI tools.
• Integrated Assessment – Integrating the measurement of academic understanding with the analysis of states, such as motivation, attention, self-regulation, and engagement, to provide holistic insights and inform social-emotional instruction.

National Investment & Impact

Federal investment in AI and assessment sets the foundation for national innovation. By fostering an open, accountable, and AI-enabled ecosystem, policymakers can ensure AI-driven education tools are effective, equitable, and aligned with economic and workforce needs.

The Study Group’s R&D roadmap will guide AI-driven transformation by:

• Establishing public-purpose public infrastructures for AI in education
• Developing interoperable, privacy-enhancing data systems
• Supporting educators with AI-driven insights to tailor instruction

A National AI Mission for Education

AI and learning science provide a once-in-a-generation opportunity to redesign education systems that prioritize equity, skill-building, and meaningful learning experiences. By investing in AI-powered assessments and educational tools, the U.S. can ensure every learner has access to the high-quality, personalized education they deserve. This roadmap positions AI as a force for innovation, inclusion, and national prosperity.

This blog series on Advancing AI, Measurement and Assessment System Innovation is curated by The Study Group, a non-profit organization. The Study Group exists to advance the best of artificial intelligence, assessment, and data practice, technology, and policy and uncover future design needs and opportunities for educational and workforce systems.