Our Commitment to Responsible AI
‍Prisms of Reality Inc. is committed to the ethical and responsible use of artificial intelligence in educational technology. This policy outlines our approach to AI integration, governance, and protection of student data and privacy. AI serves as a targeted feature within our platform, designed to provide just-in-time feedback to students to assist them in sensemaking and articulation, while our core educational content and pedagogy remain extremely human-centered.  

1. How We Use AI Our platform incorporates AI technology for the following limited and specific purposes:
• Providing just-in-time feedback to students: Students submit answers to multiple choice and short answer math and science questions. A few seconds later, our AI provides short, written feedback to the student articulating the strengths and weaknesses of their response and reasoning. 
• Classifying student mastery: Based on students’ answers, our AI will classify students in categories such as “Ready to move on”, “Nearly there”, or “Needs review”. These classifications are provided to the teacher, along with students’ responses, to further guide instruction. Teachers can override the classifications.
• AI features are clearly identified within the platform and use of them is optional. Traditional, non-AI functionality remains fully available for all users.  

2. Data Privacy and Security We maintain strict safeguards for any and all data provided to our AI:
• Student data is never used to train third-party AI models
• No PII or other sensitive data is fed into the AI
• No data fed into the AI is shared beyond our specified AI service providers
• All AI processing complies with FERPA and COPPA
• Data minimization principles ensure only necessary information is processed by AI systems
• Data protection is governed by AWS’ Bedrock Data Protection. User guide is available here.  

3. AI Technology Providers We partner with industry-leading AI providers that meet our rigorous standards for data protection:
• Primary AI Provider: We utilize foundation models built by Anthropic and accessed through AWS Bedrock   

4. Human Oversight and Control We leverage AI as an assistive tool, not as a replacement for educator judgment and expertise:
• Educators maintain full control over how and when AI features are utilized in their classrooms
• Educators retain the right to override the AI classifications at any time
• Students and educators can report concerns about AI outputs through various feedback channels
• Regular audits assess AI performance, bias, and educational effectiveness  

5. Bias Mitigation and Fairness We leverage AI in a way that inherently mitigates bias to ensure equitable educational outcomes:
• This is achieved primarily by not giving our AI access to any race, sex, gender, ethnic, or other data, as well as no identifiable information about the person submitting responses for evaluation by the AI
• Rigorous, standards-aligned training data and validation sets to minimize any risk of non-domain-specific output  

6. Content Accuracy and Quality We hold the rigor and accuracy of AI-generated content to be paramount:
• AI outputs are validated against a rubric generated and refined by Prisms employees
• The rubric is reviewable on request Users can flag inaccuracies, and we maintain response protocols for corrections
• Educators can report AI-generated content that appears inaccurate or imprecise through various feedback channels  

7. Parent and Student Rights We respect and support the rights of students and parents regarding AI use:
• Parents can request information about AI features their children interact with while using Prisms by emailing any questions to support@prismsvr.com
• Educators can utilize Prisms’ products without requiring students to interact with AI features
• Students have the right to understand when they're interacting with AI versus human-created content
• Data submitted by students to Prisms’ AI can be retrieved and audited on written request. It is also always available to the student’s teacher  

8. Governance and Continuous Improvement We evolve our AI practices along with improvements in technological capabilities and regulatory requirements:
• Regular review of AI policies by a cross-functional team
• Staying current with educational AI research, best practices, and regulatory requirements
• Soliciting feedback from educators, students, and parents to improve AI features
• Our use of AI features is reviewed and updated continuously in response to user feedback and standard product development cycles
• This policy is reviewed and updated at least twice annually   

Contact: For questions about this AI policy, please contact support@prismsvr.com
Version: 1.0