Using AI to Strengthen Authentic Assessment—and Better Instruction
Every teacher has experienced it.
The essays have been submitted.
The math problems have been solved.
Student work is waiting to be reviewed.
Somewhere within those papers are the insights that should shape tomorrow’s lesson.
Which concepts have students mastered? Where are misconceptions beginning to appear? Who is ready for enrichment, and who needs additional support?
Finding those answers takes time—and often, by the time every paper has been reviewed, the opportunity to immediately adjust instruction has passed.
That challenge was the focus of JFYNetWorks‘ MAVA 2026 session:
From Evidence to Action: Using AI to Amplify Authentic Assessment in ELA & Math.
Rather than asking teachers to change how they assess student learning, the session explored how AI can help educators make better use of the authentic assessments they already trust.
Authentic Assessment Is About More Than Grading
Whether evaluating argumentative essays in ELA or reviewing students’ mathematical reasoning through constructed responses, teachers rely on authentic assessment because it reveals far more than whether an answer is simply right or wrong.
Authentic assessment uncovers how students think, communicate, reason, and apply their learning.
The challenge isn’t collecting evidence.
It’s having enough time to analyze that evidence while it can still influence instruction.
The JFY AI Teacher Assistant was designed to help educators organize and interpret student work more efficiently—without replacing the professional judgment that only teachers can provide.
AI That Begins with the Teacher
One of the most important ideas presented during the session is that this isn’t generic AI feedback.
Everything begins with the teacher.
For an ELA assignment, teachers select or create the writing prompt, identify the reading selections, establish the rubric, and provide benchmark examples that define quality writing.
For mathematics, teachers create or select the constructed-response problem, align it to standards, define the scoring criteria, and establish the expectations students are working toward.
Only then does the AI evaluate student work.
In other words, the AI learns the teacher’s expectations before it ever reviews a student’s work.
Because the system is built around teacher-created assignments and teacher-defined rubrics, the resulting feedback reflects the expectations of the classroom—not generic responses generated from the internet.
Turning Student Work into Instructional Evidence
Once an assessment has been created, teachers can upload an entire class set of student work in a single workflow.
In ELA, the system produces individualized feedback on writing, organization, use of evidence, language conventions, and other rubric-aligned criteria.
In mathematics, it evaluates students’ reasoning, mathematical communication, problem-solving approaches, and constructed responses using the teacher’s own scoring expectations.
At the same time, teachers receive something equally valuable:
A whole-class summary.
Instead of searching for patterns one paper at a time, teachers receive an organized view of class-wide strengths, recurring misconceptions, and opportunities for reteaching while the learning is still fresh.
From Individual Feedback to Better Instruction
One of the presentation’s strongest messages was that assessment shouldn’t stop with individual student comments.
When viewed together, assessment results reveal something much larger.
Teachers can quickly identify:
- rubric criteria students understand
- concepts that require reteaching
- opportunities for small-group instruction
- patterns across the entire class
Instead of simply assigning scores, assessment becomes a practical planning tool for tomorrow’s instruction.
Measuring Growth Over Time
Learning rarely happens in a single draft.
Whether students are revising an essay or strengthening their mathematical reasoning across multiple problem-solving opportunities, growth develops over time.
The JFY Teacher Assistant helps preserve that growth by creating downloadable assessment records that allow teachers to compare student performance across drafts, units, and classrooms.
👉 Learn more about the JFY AI Writing Assessment Tool and the instructional approach behind it.
Rather than isolated feedback, educators build an ongoing record of student progress—making conferences, intervention planning, revision, and progress monitoring more purposeful and data-informed.
Teachers Remain at the Center
Throughout the presentation, one theme remained constant:
- Teachers remain in control.
- Teachers create the assignment.
- Teachers define the rubric.
- Teachers interpret the evidence.
- Teachers make the instructional decisions.
- AI simply helps them get there faster.
That distinction matters.
The goal isn’t to automate assessment.
It’s to make authentic assessment more practical—and more actionable—for today’s classrooms.
Teachers remain at the center of every instructional decision. AI simply helps bring the evidence into focus.
Looking Ahead
As schools continue exploring the role of AI in education, the most meaningful applications may not be those that replace teaching—but those that strengthen it.
Authentic assessment has always generated meaningful evidence of student learning.
The challenge has never been collecting the evidence. It’s finding the time to act on it.
When AI helps teachers uncover insights more quickly—without replacing their professional judgment—that evidence becomes something even more valuable:
Action.
Continue the Conversation
Interested in bringing these ideas into your own classroom?
Download the complete presentation, From Evidence to Action: Using AI to Amplify Authentic Assessment in ELA & Math, to explore the instructional framework, teacher-defined assessment workflows, and practical classroom examples shared during the session.
Interested in Learning More?
JFYNetWorks partners with schools and districts to help educators integrate AI thoughtfully, ethically, and instructionally—always with teachers at the center of the learning process.
👉 Schedule a complimentary AI consultation with JFYNetWorks.
