Chip Li李鸿铭
University of FloridaVIABLE LabLOGEN AI

AECT 2026 · Emerging Learning Technologies · Chicago

Don’t Write It
For Me

Class-Ready is the LOGEN AI mode for learning to write objectives. Learners draft an objective, check it against the criteria, and revise in response to an AI question. Instructors can see each draft and the learner’s reasoning.

The questionHow can AI feedback help teacher candidates write their own learning objectives?
The designLearners check their drafts first. AI asks a question to guide revision.
What we sawA learner wrote clearer objectives even as agreement with AI fell.

Two ways to use LOGEN AI

Studio gives educators a starting draft for a lesson. Class-Ready gives teacher candidates practice deciding what students should do and how to tell if they succeeded.

2025 LOGEN AI generates and reviews, and the educator decides.2026 Class-Ready adds a classroom workflow where learners write every draft and AI responds with a question.

LOGEN AI at AECT 2025 and today

We introduced the Classic objective generator at AECT 2025. It grew out of our ARCHED framework. The current LOGEN AI offers that drafting and review workflow in Studio, alongside Class-Ready for courses that teach objective writing.

Classic LOGEN AI with three columns for generator settings, previews, and saved objectives
LOGEN AI ClassicThe Classic objective generator from the 2025 design
Current LOGEN AI home page with an example lesson and an entry point to the Studio
LOGEN AI todayStudio for educators, Class-Ready for learners

What the AI does in each mode

Both modes use AI to review Bloom level, SMART criteria, and fit for the learners. That is the idea behind LLM-as-a-judge. Compare what the AI does after reviewing the same draft.

A draft objective“Students will do a worksheet on fractions using fraction strips.”

The judge, in both modes
ApplySMART

Completing a worksheet is an activity, and the fraction skill or evidence being assessed is not identified.

Ask yourself

What fraction task would students complete with the strips, and what response would show their reasoning?

No sentence to copy. The next version of this objective has to come from the learner, and the server refuses to write it for them.

LOGEN StudioClass-Ready
Generate and refineYesNever in class work, enforced on the server
ReviewRewrites with one-click applyJudgments, the main gap, one guiding question
Who sees the workThe educatorThe learner and the class instructor
EvidenceThe final objectivesThe process that produced them

How a learner uses Class-Ready

The learner checks each draft before seeing feedback. They can revise and repeat the process until they are ready to submit for their instructor to review.

LearnerAI judgeInstructorWriteCheck yourselfAsks a questionReviseSubmitGrade returnedAI readRead the recordGrade & release
Learner

Judge first

Before any feedback, the learner picks the Bloom level and the S·M·A·R·T criteria each ready objective meets. The judge never sees this check.

Self-check dialog asking which level the task demands and which criteria it meets

From Proposal to Production

Our AECT proposal described a stand-alone prototype, and its figures show that version. Three rounds of development with experts followed, and we shipped the result as a public, production system inside LOGEN.

In the proposalLive today
A stand-alone app for pre-service teachersA mode inside LOGEN, using the same judge. Learner status comes from joining a class, and anyone can create one.
AI offers concrete suggestions for revisionAI names the main gap and asks one question. The server refuses to write objectives in class work.
The learner submits, then AI evaluatesThe learner judges level and S·M·A·R·T first, before every round. The judge never sees that check.
No instructor in the pictureInstructors create classes, read the process, grade with a rubric, and release grades when ready.
Exactly two Common Core standardsThe instructor sets them, learners choose them, or the assignment goes without. Custom standards are welcome.
Formatted text to copy into CanvasA receipt link that opens the submission and how it took shape.

Try It

Class-Ready is public and free. Any account can learn in one class and teach another.

As a learner

Open Class-Ready and choose Try the open demo class. No code needed. Sign in with Google, write an objective, check it yourself, and read the feedback.

As an instructor

Sign in and open Classes in the left sidebar. Create a class, set an assignment, and share the code. Try it yourself walks you through your own assignment without counting toward the class.

What a learner sees

Screens below come from a test class.

What an instructor sees

How We Tested It

Three recruited participants tried the learner and instructor workflows using the live AI review. They thought aloud as they worked, rated each task, and talked through their decisions in an interview.

3recruited participants
33tasks, from first look to end of term
26objective reviews by the real model
515screenshots, with 3,278 logged events

Participants completed 25 tasks as expected, 2 with extra effort, and 6 in part. We compared their accounts with the recorded drafts and feedback.

One question made success measurable

In round 2, the AI asked how Jordan would know students could use all three fraction comparison symbols, including equals. In round 3, Jordan added equal pairs and set a threshold of eight correct answers out of ten. Follow objective #4 below to see how the draft changed.

#2 Students will learn about ordering fractions from least to greatest.

You Understand · Review UnderstandSMART

The objective names a topic rather than one observable learner action and its evidence.

Ask yourselfWhat could a learner do or show that would demonstrate they can order fractions?

#3 Students will do a worksheet on fractions using fraction strips.

You Apply · Review ApplySMART

Completing a worksheet is an activity, and the fraction skill or evidence being assessed is not identified.

Ask yourselfWhat fraction task would students complete with the strips, and what response would show their reasoning?

Agreement with AI
How often the learner and AI made the same checks
Specific and Measurable
How often AI marked these two criteria as met
0%50%100%Round 1: 0% of Specific and Measurable grantedRound 1: 58% agreement0%58%Round 1Round 2: 25% of Specific and Measurable grantedRound 2: 13% agreement25%13%Round 2Round 3: 63% of Specific and Measurable grantedRound 3: 83% agreement63%83%Round 3
Agreement and criteria granted by round
RoundAgreementSpecific and Measurable granted
158%0%
213%25%
383%63%

The AI never writes or rewrites the objective for the student, and the student judges their own work before seeing feedback.

Renee, instructional designer

Clearer drafts can come with lower agreement

In round 2, Jordan replaced phrases like “understand” with actions students could demonstrate. The AI marked more Specific and Measurable criteria as met. Yet only 13% of Jordan’s self-check decisions matched the AI, down from 58%. In the activity below, tick Achievable, Relevant, and Time-bound to see what happens to that figure.

#1

Students will compare two fractions with different denominators and say which one is bigger.

You Analyze · Review Apply

YouReviewSMART
#2

Students will put fractions in order from least to greatest and show how they got it.

You Apply · Review Apply

YouReviewSMART
#3

Students will use fraction strips to compare two fractions on a worksheet and write why one is bigger.

You Understand · Review Apply

YouReviewSMART
#4

Students will use the symbols >, =, and < to show which fraction is bigger.

You Remember · Review Apply

YouReviewSMART

How often the review granted each criterion, across all 26 reviews

S21
M13
A25
R26
T26

Those three boxes raise agreement to 63% while the draft stays the same. The AI marked them in 77 of 78 checks, and the default rubric gave agreement 2 out of 10 points. A rubric that asks learners to explain their checks would assess the skill they are practicing.

That 13% was the one moment I wanted to quit, and it wasn’t even measuring my writing.

Jordan, pre-service teacher

The AI summary missed changes in the final draft

During the learner task, the participant’s round 3 note asked, “I’m not sure if the others need a number too.” Then the participant added one to each objective and submitted, with no further review. When a participant in the instructor role asked for an AI read, it said those objectives lacked a threshold. Pick an objective to compare.

Round 3, the last review

Students will compare two fractions with different denominators, say which one is bigger, and draw a picture that shows how they decided.

SMARTMeasurable not met
The AI read of the submission

“Review #1–#3 for measurable success criteria: each specifies an observable task but lacks a clear accuracy threshold.”

Rubric rowAI suggestedParticipant rating, after reading the record
Final objectives 4 ptsDeveloping 2.7Proficient 4
Use of feedback 2 ptsDeveloping 1.3Proficient 2
Judging their own work 1 ptsBeginning 0.3Developing 0.7
Reasoning about changes 3 ptsDeveloping 2Proficient 3
Total of 106.39.7

The AI summary received the final wording alongside earlier reviews without a marker for the revision. The participant acting as instructor could see the success thresholds in the submitted draft and score the work from that record.

Right now the tool’s own summary penalizes the student who goes on thinking without paying for another round, and that undermines the entire promise of seeing the process.

Renee, instructional designer

What instructors can check against the record

“The full class of 24”

That is the number of fourth graders in the assignment brief. The test class had two learner participants.

Two learners became one

The class read described one learner as “essentially unchanged and stuck.” The unchanged work and the stuck work belonged to different participants.

“Stuck” on the most revised work

The overview flagged the participant who changed the most as stuck, because the label also relies on the last reviewed version.

“59% pasted” in amber

Someone who retypes AI text shows 0% pasted and a green tag. The caution sat in a tooltip, and the risk falls on second-language writers who draft with help.

In Their Words: Participant Interviews

All three participants wanted to keep the self-check and the questions. Renee and Min-seo trusted the draft history most because they could see what a learner had changed. Jordan found the written feedback useful for deciding what to revise.

Recruited participant

Jordan, pre-service teacher

A nervous, time-poor novice who writes “Students will understand” and has lost points without knowing why. Learner in one class.

Tasks succeeded
7 of 11
Mean SEQ
5.1
Clarity
5/7
Ease
5/7
Helpfulness
6/7
SUS
62.5

Next term Would recommend it to classmates for a class that assigns it. Wouldn’t use it alone while AI is paused for student teaching.

That’s the first time anybody explained why ‘do a worksheet’ isn’t an objective, and it did it by asking me a question.

Jordan, pre-service teacher

It told me the one thing I actually keep getting wrong, and it’s the first feedback I could carry to my next lesson plan.

Jordan, pre-service teacher

I finally get what “measurable” means, and I’d use that on my next lesson plan, but some of the numbers didn’t help.

Jordan, pre-service teacher

What We Take From This

Keep the question visible during revision

A specific question gives the learner something to act on in the next draft.

Formative assessment · Black & Wiliam, 1998

What changes Show the question beside the draft as the learner revises.

Link each AI claim to a draft

The draft history let the instructor spot where the AI summary described an earlier version.

Feedback literacy · Carless & Boud, 2018

What changes Mark revisions made after the last AI review and link the summary to the version it describes.

Give credit for reasoning

When learners explain why a criterion applies, instructors can see how they judged the draft.

Self-regulated learning · Nicol & Macfarlane-Dick, 2006

What changes Let learners mark “I disagree because…” and credit the reasoning. Show the rubric from the start.

What Comes Next

Make feedback easier to checkShow the rubric up front and link AI summaries to the drafts they describe.
Expand the participant studyUse the same tasks and interview guide with a larger sample in teacher-education courses.
Explore other teaching tasksTry question-based feedback for assessment design and lesson planning.

The questions we most want answered:

  • Does the 13% moment discourage learners, or do they shrug it off?
  • Does judgment grow over rounds, and does it transfer to objectives written without the tool?
  • How does knowing an instructor can see every draft change how people write?
  • How does a busy instructor with thirty learners read an amber “pasted” flag?
  • How do second-language writers experience the paste note?

Conference Talk

See how a teacher candidate used an AI question to make a learning objective more specific and measurable.

AECT 2026 talk

Don’t Write It For Me

References

  • Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education: Principles, Policy & Practice, 5(1), 7–74.
  • Carless, D., & Boud, D. (2018). The development of student feedback literacy: Enabling uptake of feedback. Assessment & Evaluation in Higher Education, 43(8), 1315–1325.
  • Li, H., Fang, Y., Zhang, S., Lee, S. M., Wang, Y., Trexler, M., & Botelho, A. F. (2025). ARCHED: A human-centered framework for transparent, responsible, and collaborative AI-assisted instructional design. Proceedings of Machine Learning Research, 273.
  • Nicol, D. J., & Macfarlane-Dick, D. (2006). Formative assessment and self-regulated learning: A model and seven principles of good feedback practice. Studies in Higher Education, 31(2), 199–218.
  • Zheng, L., Chiang, W.-L., Sheng, Y., Zhuang, S., Wu, Z., Zhuang, Y., … Stoica, I. (2023). Judging LLM-as-a-judge with MT-Bench and Chatbot Arena. Advances in Neural Information Processing Systems, 36, 46595–46623.