As artificial intelligence (AI) transforms education and workforce assessment, one question rises above the rest: how can we trust the systems that will shape opportunity for billions of learners and workers? At the Center for Responsible AI in Learning and Assessment, our answer is trust begins with measurement science.
Why measurement science matters
Measurement science is the backbone of fair, valid and reliable assessment. For decades, ETS has set the global standard for psychometrics, ensuring that every test, score and evaluation is grounded in rigorous evidence and advancing foundational research that has pushed the boundaries. Now, as AI enters the picture, this expertise is more essential than ever.
AI-powered assessment tools promise efficiency and personalization, but they also introduce new risks: bias, “black box” algorithms, and the potential to exclude rather than empower. Without robust measurement science, these risks can go unchecked. With it, we can build systems that are not only innovative, but also defensible and equitable.
ETS’s unique approach
At the Center for Responsible AI in Learning and Assessment, measurement science is woven into every aspect of our work:
- Independent evaluations: We conduct rigorous, third-party reviews of AI systems to ensure they meet the highest standards for fairness and validity.
- Standardized benchmarks: Our criteria and metrics — rooted in decades of research — enable organizations to objectively evaluate performance, reliability and equity.
- Purpose-built auditing tools: We provide actionable resources for education and workforce leaders to monitor, assess and improve AI-enabled assessments.
- Evidence-based guidance: Our recommendations are grounded in real-world data, transparent methodologies and reproducible results.
- Foundational research: We develop novel frameworks, methods and capabilities that aim to reshape the future of measurement science and human learning.
From research to real-world impact
What does this mean for educators, policymakers and communities?
- Defensible decisions: Leaders can implement AI solutions with confidence, knowing they are backed by independent evidence.
- Equitable outcomes: Every learner’s achievements are measured fairly, opening doors to opportunity and mobility.
- Transparent systems: Stakeholders can see how decisions are made, fostering trust and accountability.
Building public-interest infrastructure
The Center’s commitment to measurement science isn’t just about technical rigor — it’s about building infrastructure that serves the public good and advancing research that redefines what’s possible. By conducting foundational research and translating it into practical tools, policy frameworks and training programs, we help organizations reduce risk, ensure fairness and scale impact.
As we launch, our goal is simple: to make AI in assessment trustworthy, transparent and equitable for all. With ETS’s measurement science as our foundation, we’re turning research into infrastructure that communities can rely on.