At this moment in time, machine learning has evolved where most of ed-tech is operating in a post-generative space, when artificial intelligence (AI) has already fundamentally reshaped the landscape of educational tools. This means that not only are students and educators learning in a technology-mediated way, but where tech-free explorations show up are vastly different. For example, a child’s first encounter with self-guided reading of a novel or chapter-book might be on a screen rather than a printed page and with one click, they may be able to generate different versions of that text that increase relevance or provide a more personalized approach to how they are understanding foundational literacy skills. Even more, assessment is now embedded in such a way that it has become embedded and indistinguishable from discrete learning tasks.
Educators are also utilizing a more AI-informed approach for how they prepare instruction, even as we are still trying to optimize how to leverage these capabilities to retain the flow that educators hope to see during a lesson arc. This expands to beyond the classroom, for the use of frontier models (i.e., Anthropic, Google, and Open AI) have made it even easier to integrate AI into your daily decision-making, such as through meal planning, travel, and daily prompting for mental health, whether you are 13 or 98. This means that a third horizon most certainly will include the need to grapple with how AI is impacting education, not if AI will impact education.
Even in the race to include generative capabilities in our tools, a call remains, “How might we truly capture the full spectrum of young people’s growth and learning in ways that prepare them for this transformed landscape?” Assessment for Good, a research and development program supported by the Advanced Education Research and Development Fund (AERDF), in partnership with Knowledgeworks, asked of the field “What are the Futures of Assessment?” In this interactive report contributors forecasted To begin to tackle this, we must also question the role of “the test.” A time-honored ritual of bubbled answer sheets and standardized metrics, that even now in its computerized form, captures only a narrow slice of human capability. With the time assessment takes, including the preparation by teachers, that precious instructional time is being taken by a practice that is not telling us all we should know about student experience and learner development. It is akin to trying to understand the ocean by examining a single drop of water, standardized testing reduces the rich complexity of learning to isolated data points that only sample our academic standards, which tells us very little about each learner, let alone if we’ve truly been accountable to them as facilitators of their future (insert reference). As we prepare young people to flourish, despite the Age of AI, we need assessment approaches that mirror the multifaceted, dynamic, and deeply human capabilities they’ll need to thrive.
The Limitations of Testing in an AI Age
The irony between AI and assessment is stark. As AI becomes increasingly capable of performing routine, replicable tasks, such as quickly finding predictable patterns and connections amongst data, our assessment procedures continue to ask learners to perform similarly narrow tasks, such as recall information and execute procedures to show their skill mastery. What AI cannot do, and what will become increasingly valuable as we try to capture more dynamic learning, are the uniquely human capacities that come from experience — creativity, emotional intelligence, cultural understanding, ethical reasoning, and the ability to navigate ambiguity and build meaningful relationships. Thus, our assessment practices are both enhanced and held back by the promise of AI if we are not careful. As a field, absent this capability, we have yet to capture growth in such a way to promote learning that is rich, real-world relevant, and connected to a learner’s experience.
A choice remains, to stay in the limitations of traditional testing, which endangers a future of learning that perpetuates what we name as a “deficit view” of learners, particularly those furthest from opportunity, such as learners situated in poverty and learners with disabilities. Our assessment practices often ask students to prove their worth against narrow, biased standards rather than revealing their unique brilliance and potential.
Asset-Based Assessment: A Path Forward
The concept of asset-based formative assessment offers a compelling alternative. Rather than focusing on what students lack, this approach seeks to unveil strengths and build upon them. Explored through the Futures of Assessment forecast, this approach to assessment, when implemented thoughtfully, can create conditions for capturing evidence that reflects learning that is connected, dynamic, and reflects how learning occurs, naturally. In our call for a new era of assessment, we must think beyond core content knowledge and imagine an experience that brings learning to life.
- Feedback and Dialogue: Assessment that includes ongoing conversations among learners, educators, and caregivers—a continuous exchange that informs all parties simultaneously.
- Customized Support: Recognition that each learner’s journey has its own rhythms, responding to their unique needs, interests, and goals while honoring the fuller context of their lives.
- Agency: Developing learners’ capacity to act on the world, expanding how, when, and where they engage in discovery and assessment.
- Expanding beyond the Content: Attending to holistic learning, fostering development, connection, and engagement to learning environments.
- Life-long learning: Creating pathways that extend learning opportunities into communities, breaking down artificial barriers between school and life.
Capturing Evidence in New Ways
So how might we actually capture evidence of growth and learning beyond the test? This requires a reimagination of what we measure and how that relates to student learning. Think about this dichotomy within portfolio assessment: A student brings together all of their assignments into a folder in their cloud-based classroom account. They upload all of their assignments and then, an embedded LLM summarizes the assignment for each teacher, automating the teacher review of materials and suggesting a grade. Conversely, the same capability could be utilized in a more connective way. That same portfolio assessment could be a chat-based interface that helps learners document their own learning journeys through self-reflection and peer assessment on classroom assignments, building crucial metacognitive skills. As students upload an assignment, they are asked to examine patterns in their individual or group-mediated learning. As the quarter continues, they develop the self-awareness needed to direct their own growth as they explore connections across their assignments.
These twists on capability integration will be critical for us to not only monitor, but to engineer towards a future that is not just plausible, but preferred, to position AI as a facilitator of learning that is more “customizable and transparent” rather than “prescriptive and opaque.” The plausible goal is a re-imagined future that automates assessment. However, a preferred future could be the opportunity for us to understand learner development, to understand the contexts across which that development occurs, and capture that learning in such a way that allows trusted adults to see their brilliance more fully and make decisions that are rooted in facilitating that learner’s maximized potential.
When we think about the technological surprise that we could garner under this era of fast-moving innovation, we should be courageous to move farther than just an agent that summarizes, to generative and traditional AI applications that could help us recognize patterns across diverse forms of evidence—from project portfolios to community service, from creative expressions to collaborative problem-solving. Breakthroughs in this vein can help translate rich, multifaceted learning experiences into meaningful insights, without reducing them to simple scores of learner status and proficiency—preserving instead the complexity that makes each learner’s journey meaningful.
A New Era Imperative: Building Trust and Community
This means that the definition of assessment and the role of “the test,” has to move beyond sitting in front of a computer within four school walls. It has to involve families and communities as genuine partners in the process of assessment in service of learning, recognizing the funds of knowledge that exist outside school walls. It means creating the opportunity for young learners to see themselves as part of a larger narrative of growth and contribution.
By moving beyond the test, we move toward an assessment and learning process that honors the full humanity of our learners and prepares them for a future where their uniquely human capabilities—creativity, empathy, wisdom, and moral reasoning—will be their greatest assets expanding our understanding of what constitutes core content knowledge across the school-age lifespan. Even more, they likely become the best levers in powering learning that exceeds beyond what we imagined as the field absorbed and ran with Bloom’s two sigma discoveries. In the age of AI, reimagining assessment isn’t a nice to have; it’s an essential step for finally supporting individual flourishing in a customized way towards maximized lifelong success and opportunity.
1 https://www.the74million.org/article/drawing-on-video-games-educators-land-on-unlikely-idea-playful-assessment/
2 https://news.gallup.com/poll/691967/three-teachers-weekly-saving-six-weeks-year.aspx
3 https://www.timesunion.com/news/article/ai-generated-lesson-plans-fall-short-on-inspiring-21230485.php
4 https://ihpi.umich.edu/national-poll-healthy-aging/national-findings/how-older-adults-use-and-think-about-ai
5 https://futureofassessment.org/
6 https://www.k12dive.com/news/how-do-you-get-the-right-amount-of-testing/448207/
7 https://www.k12dive.com/news/how-do-you-get-the-right-amount-of-testing/448207/
8 https://ascd.org/el/articles/the-future-of-assessment-is-now
9 https://knowledgeworks.org/resources/tool-exploring-plausible-probable-possible-preferred-futures/
10 https://onlinelibrary.wiley.com/doi/10.1111/emip.12370