Our AI sub-processor, OpenAI, is held to the same high data privacy and security standards as all other sub-processors. They are GDPR compliant and are certified with leading industry privacy and security standards including ISO27001, SOC 2 and SOC 3. Find OpenAI’s Privacy Policy here, and their Security Portal here for more information.

The generate contextual and constructive feedback suggestions for teachers to share with students, our AI Feedback Assistant will include the following data points in your prompt:

Any data points that have not been provided will not be included (such as a class that does not have a year group identified). No PII (personally identified information) from either the teacher or the student account is provided to OpenAI. The only exception is if a teacher were to include PII in their initial written feedback, which we advise against.

Data input into our AI feature as prompts will be stored by OpenAI for a maximum of 30 days in order to monitor for abusive language before being fully deleted. As per the terms of our subscription agreement with OpenAI, data provided to OpenAI by Showbie in the form of a prompt is not used for further training of the OpenAI model and is not retained or used by OpenAI beyond the standard 30 days of content monitoring noted above.

Only the answer content and question context are included, and no student identifiers of any kind are sent to OpenAI.

This means:

The only exception is if a student includes personal information inside a written short-answer response, which we advise against. These responses are still sent without any linked identifiers, and OpenAI processes them only transiently to generate insights.

No. Showbie’s AI features do not perform grading or automated decision-making about students.

All AI features are designed strictly to support teacher-led instruction and remain fully aligned with emerging AI-in-education regulations (including restrictions on automated grading and profiling).

Yes. While AI systems deployed within the education field are considered high-risk, Showbie's use of AI is limited to the following acceptable intended use:

Showbie is compliant with Chapters I and II of the EU AI Act which will be enforced as of February 2nd, 2025.

No. We plan to support a continuous opt-out option for the use of AI, meaning that you or your school/district can select to turn off access to AI features if you don't wish to use them. Additionally, the AI feature will be included in Pro subscriptions so access will be limited to only Pro teachers, and there are no limitations in your use of Showbie if you choose not to engage the AI feature.

That being said, we have taken the opportunity while developing our AI feature to further develop our Gradebook feature, as well as to add optional contextual information to classes and assignments. These updates to the Gradebook and class and assignment info are available to all users and will improve your experience when grading student work or providing formal feedback whether or not you use features such as AI Feedback Assistant or AI Insights.

We recognize that inaccuracies and biases still exist within the capabilities of generative AI and the content it's trained on, which is why all AI-generated output in Showbie (including AI Feedback suggestions and AI Insights) must be reviewed by the teacher before being used, shared, or relied upon for instruction. Additionally, we encourage users to flag AI-generated content that is inaccurate or inappropriate so that we can work to improve future results.

While there are no technical limitations in place, you are discouraged from including personal data (such as names, email addresses, birthdays, etc.) in the feedback that will be used within your prompts, as well as abusive or derogatory language or profanities. OpenAI will monitor prompt inputs for abusive language for up to 30 days after the prompts are received.This applies both to teacher-written feedback used in AI Feedback Assistant and to free-text student responses used in AI Insights.

No, our AI feedback feature is solely teacher-facing and will not be visible to students. The extent to which students will experience our AI features is that AI-generated content will be visible to students following review and approval by their teacher.

No, there will be no student data shared with OpenAI by our AI feedback feature. The content generated by our feedback feature will also require review and approval by the teacher before becoming visible to student users, with options to edit or discard the AI-generated content. As with all of our sub-processors, only the limited, necessary data required to provide the feature will be collected and shared to OpenAI.

The AI Feedback feature relies on initial feedback provided by the teacher, in combination with class and assignment description information, to generate a prompt for the AI tool to analyze. Internally, our AI Feedback tool is designed to focus on grade-appropriate suggestions to improve the feedback provided by the teacher, as well as incorporate suggestions of next steps for the student. The associated student account and student submissions will not be included in the prompt or visible to the AI tool, which is why we have decide to offer the opportunity to provide further description information about classes and assignments to help ensure feedback is contextually appropriate. Placeholder text is used where needed to allow the teacher to provide further details, personalized to the student.

In a similar way, our AI Insights feature also avoids accessing or processing identifiable student information. To generate class level insights, the tool receives only anonymised response data from the completed assessment, for example whether students answered a question correctly, question metadata, or the text of a short answer response without any associated student identifiers. The AI does not receive names, emails, user IDs, profile information, or anything that can be traced back to a specific student. Insights are based solely on aggregated patterns in the anonymised responses, and the output is visible only to teachers.