Question: What’s new about the updated AI feedback model, and how does it improve on Version 1?
​Answer: Version 2 uses sample submissions (real student writing curated by the Writable Team) to “train” the AI to recognize the characteristics of strong, average, or weak responses. This results in more accurate scores and feedback better aligned with your specific rubric criteria. Click here to learn more about the changes.

Question: Will using the new model require extra setup or PD (professional development)?
​Answer: No. If you’re comfortable using Writable's AI feedback with Version 1, you’ll find Version 2 works similarly, just with enhanced accuracy. However, you can access short tutorials and tips to help you understand how to use Writable's AI grading supports.

Question: Is Version 1 still an option for ongoing or existing assignments?
​Answer: Absolutely. Version 1 remains available for consistency in tracking student progress or maintaining yearlong benchmarks. You can switch models in the “Edit Assignment” settings. Click here for step-by-step guidance.

Question: Does Version 2 mean I’m no longer needed in the feedback process?
​Answer: No. AI feedback is a support tool, not a replacement for teacher expertise. It handles initial scoring and provides suggestions. You should always review, adjust, or override the AI’s scores and feedback as you see fit.

Question: What prevents bias or inconsistent scoring across different classrooms or assignments?
​Answer: By using curated sample submissions as benchmarks, Version 2 develops a clearer understanding of each rubric level, reducing the chance of over- or under-scoring. The result is more consistent performance across assignments and classrooms.

Question: Does this model work well with various writing genres and grade levels?
​Answer: Yes. Version 2 is designed to adapt to diverse prompts and genres, as long as your rubric aligns with those writing types.

Question: Can I add personalized comments or override the AI’s suggestions?
​Answer: Yes. The AI’s feedback is a starting point. You’re always in control of final scores and any additional comments you feel are necessary to support individual students.

Question: Sometimes I interpret a student’s writing differently—can I adjust the score?
​Answer: Of course. You can manually edit scores and comments. The AI is designed to assist you, but ultimately, teacher discretion is key to providing the best feedback for each student. If you find that the AI feedback seems very off in it's assessment of the writing please share those examples with us by emailing support@writable.com.

Question: I’d like to add my own sample submissions that the AI uses for calibration. Is that possible?

Answer: Not at this time. The sample submissions are curated “behind the scenes” from writing exemplars in each writing genre. Because they’re part of the Writable AI’s internal calibration process teachers don’t have direct access to them. Rest assured, these samples are carefully selected to ensure the AI provides reliable, consistent feedback across all student writing.

If you have any questions about our new AI scoring model please contact our team at support@writable.com. Including assignment details and relevant screenshots will help us investigate issues more efficiently.