1) Create opportunities for teachers to see these shifts in action; data driven, actionable and authentic example that the shift works.

2) We should be open and vulnerable to lead the shift and be raw about sharing the positives and areas for improvements.

Raising awareness

Deepen their understanding

Phase in the implementation

How can AfL be embedded across CZ classrooms?

• having some protocols/structures to solve problems & their discourse

How can we involve students in assessment process?

• peer-critique or peer tutoring

• SEE 3 before ME

• Teacher-of-the-Day: getting students to be in front leading the discussion

• use of videos (for PE lessons or word problems) for students to critique

How might you leverage on Tech or Gen AI to augment assessment process?

• be intentionally and conscious about the IOs to feed the tool when using the Data Assistance

• Using prompt engineering to evaluate students' work

(Gerardo, Norisah, Lew Cheng, Emelyne)

How can AfL become more systematically embedded across CZPS classrooms?

1. Tap on department resources.

2. Adopt a common set of school-wide resources that are prevalently used across all subjects. To also set expectations and culture (e.g. exit slip foirmat). However, 2 sets of concerns: May not benefit all groups of students equally (e.g. due to self reporting nature); May not be contextualised to subject/ task specifically

3. To have a common structure for all lessons (E.g. End lesson 5 min earlier, to reflect on learning and understanding)

How to involve students in assessment process?

1. Construction of success criteria so that students have ownership and have a stake in what they should achieve.

2. Self assessment

3. Peer assessment and peer feedback

4. Use of checklists and rubrics

How might you leverage on technology or Gen AI to augment assessment processes?

1. Generate rubrics for self/ peer feedback

2. Create AfL task

3. Create checklist for assessment

4. Aggregate student responses for AfL for teachers to surface common misconceptions/ gaps

5. Provide instant feedback directly to students

To infuse FA into SOWs to be more intentional.

Open classroom, STP as a guide

Whole school structure (AFL) in place such as

1)Fastkits.

2)anchor charts for success criteria

✅ Accuracy

• Conceptual Error (Left Side):

  • The explanation is correct. The student placed the 4 in the wrong place value, indicating a misunderstanding of how division works in terms of place value.

  • The note “The 4 should have gone in the tens place…” is accurate and shows understanding of how 24 ÷ 6 = 4 should be interpreted.

• Computational Error (Right Side):

  • Also accurate. The student correctly identifies that the person understood the process but made a multiplication error: 3 × 6 = 24 was correct, but they put 3 as the answer for 24 ÷ 6, likely misplacing their thinking.

✏️ Clarity

• The color coding and handwriting are neat and help in distinguishing the two types of errors.

• The step-by-step breakdown using arrows and explanations makes it easy to follow.

• The phrase “Since we are dividing 24 ÷ 6 and not 2 ÷ 6” is clear and highlights the misunderstanding.

💡 Suggestions for Improvement

1. Use Math Vocabulary Consistently:

   • Add terms like dividend, divisor, and quotient where possible. For example: “Since 24 is the dividend and 6 is the divisor...”

2. Next Steps for the Student:

   • Practice identifying types of errors in different problems (e.g., subtraction, multiplication, multi-step word problems).

   • Justify correct solutions: After identifying the error, explain what the correct process looks like, step by step.

   • Peer review: Try finding and analyzing a classmate’s errors — this strengthens understanding of both concepts and common mistakes.

⭐ Overall Feedback:

This is a strong explanation with accurate distinctions between conceptual and computational errors. The student demonstrates understanding of both types and communicates it well. Next, they should try applying this analysis to a variety of math problems and incorporate mathematical language more consistently.

My concern is teachers over relying on AI to generate feedback. Teachers should have a first hand to provide feedback. Context is extremely important in the feedback cycle. Feed Up is different for each class depending on the readiness of students. Feed Back should be done at students' language and not just from Gen AI.

Feed Forward is something teachers should consider implementing it meaningfully.

We can introduce this as a way for teachers to provide feedback from group submissions.

If feedback is given with clear and actionable steps, students are then clear on what they can do to act on them.

Additionally it also helps teachers to refine and confirm their own understanding of what a good or acceptable answer would look like. It is both a feedback for students as well as teachers.

Although AI is smart and highly functional, it is important for teachers to provide constructive and understandable feedback to young learners inorder to feedup and feed forward.

This will also provide the teacher with the data to better customize their lessons to pre-empt misconceptions of students.

If students are able to key in the prompts effectively, the students would then be able to try and work on the next draft, using the feedback given and compare what it looked like before and after. Them being able to then explain to a peer in a discussion why their next draft is better will help to internalise their learning. The language used is quite easy to understand with clear examples and suggestions given. If they were to key in their age, I'm sure the AI will be able to adjust their language settings even further to suit their age.

It can provides feedback to students in a neutral and unbiased manner.

For teachers, it serves as a valuable aid in clarifying uncertainties and offering additional insights we may overlook. We sometimes miss subtle patterns or responses.

I think it is important for us also to be mindful in the things that we feed to GenAI to not compromise sensitive information.

Feedback should be clear, concise, and focused on actionable steps to help students better understand their learning gaps and take meaningful steps toward improvement.

this multi-source feedback may provide a more systematic step-by-step approach to allow students to review and improve their work.

Teacher should be mindful and also ensure that the hybrid feedback is logically and systematically sequent or include additional explanation so that students will not be confused.

The hybrid approach to providing feedback with the assistance of a Gen AI tool allows us to have a deeper understanding of the misconceptions our students have. Most of us are able to look at the mistakes and discern how to clarify the misconception but Gen AI gives a more in-depth view of the possible reasons for the student making the error. This may be useful for certain kinds of misconceptions.

It may be helpful to use AI to glean an idea on how feedback can be provided to students - e.g. structure, angles to look at. However, the teachers' lens is still important. Good to add on and supplement, make more personalised.

Question 1: Students will be able to move forward as the feedback should be more timely, specific and personal.

Question 2: With the crafted AI feedback, it will make teachers to understand better on how important timely, focussed and personal feedback will be. The crafted AI feedback can be a model (although may not be 100%) for teachers to use.

Question 3: Subject based - level teachers sharing

1) a multi-feedback loop enables student agency and teacher to address different aspects of the students' learning needs.

2) Enhancing teacher clarity - 2 key points here: enhancing and clarity- we can approach this from the perspective of deepening teacher's assessment and understanding in implementing feedback, feedforward and feed-UP

3) at the subject level, to select an area of focus for e.g. English - component base such as Oral

1) With richer feedback from multi-source, students have more to work with in their endeavour to improve. This may increase their motivation to act on the feedback to improve. The immediacy of feedback could also enthuse them to implement their improvements straight away to get 'better' feedback.

2) Multi perspectives and methods of expression could give both students and teacher greater insights into the issue. Also, reduces load on teachers.

3) Could be a PLT.

Generative AI also provided useful suggestions on how their artwork could be improved. These suggestions can be used in class by asking pupils to choose one or two ideas they’d like to try and explain why. This encourages critical thinking.

It also helps make teacher feedback clearer and supports pupils in making thoughtful, informed decisions about their work.

This activity can be done at the end of an art unit, as part of a class reflection.

Teachers will be able to better structure their feedback, and possibly use GenAI to suggest more targeted support for the students.

Possibly to trial it at for a subject and at specific level to get teachers to explore its effectiveness and limitations before considering implementation across all levels.