Our position
SmartPAL believes AI should expand human capacity in education without displacing human judgment, meaningful learning, or human relationships.
Education should not use AI merely to produce faster answers, automate instructors, or optimize engagement. It should use AI to help students learn more deeply, help educators notice where support is needed sooner, and help institutions make high-quality education more accessible.
SmartPAL’s role is to provide a course-grounded, human-led learning intelligence layer alongside the learning management system. Students should receive timely, contextual support, while instructors retain responsibility for teaching, judgment, relationships, and consequential decisions.
How this shows up in SmartPAL
These commitments shape how we design and deploy the product:
- A 24/7 AI companion grounded in course materials, assignments, announcements, and deadlines—not the open internet.
- Student support for course questions, organization, reminders, goals, reflections, and progress.
- Instructor dashboards and weekly summaries that surface engagement patterns, common questions, reflections, and opportunities for timely support.
- Canvas-based access that fits the workflows students and instructors already use.
- Learning activities that support metacognition, self-regulated learning, and critical thinking.
- Privacy, role-based access, institutional controls, and protections for unpublished course materials and unreleased grades.
- An equity approach that favors universal access within participating courses, rather than labeling or charging individual students who need help.
The through-line is simple: AI should make human teaching and learning more effective, not make them less human.
Principles
1. Learning comes before completion
SmartPAL should help students understand, practice, reflect, and make progress. It should not become a shortcut that completes academic work in place of the student.
We favor:
- Explanations, hints, questions, examples, and feedback
- Scaffolding that gradually gives students more independence
- Reflection and metacognition
- Productive struggle before immediate answers
- Prompts that ask students to evaluate and verify AI output
Success is not measured only by how quickly a student receives an answer. A better question is whether the interaction improves understanding, judgment, confidence, and independence.
2. Humans remain responsible for consequential decisions
AI may assist educators, but it should not independently determine grades, academic standing, discipline, admissions, accommodations, or other outcomes that materially affect a student’s future.
Instructors and institutions remain responsible for:
- Interpreting student circumstances
- Making high-stakes decisions
- Providing meaningful feedback and mentorship
- Deciding when an AI recommendation is appropriate
- Determining how SmartPAL is used in a particular course
SmartPAL should make human review, uncertainty, and escalation visible rather than implying that an automated recommendation is final.
3. Course context is more valuable than generic intelligence
Educational AI should be relevant to the actual course: its syllabus, concepts, assignments, deadlines, policies, and learning goals.
SmartPAL should remain bounded by authorized course context and clearly signal when it does not know the answer. Where appropriate, it should identify supporting course materials, distinguish facts from suggestions, and direct students to instructors when a question requires human judgment.
SmartPAL is not an LMS replacement and is not meant to be an unrestricted general-purpose chatbot. Its value comes from being useful inside the learning environment where students and instructors already work.
4. Early support should not become student surveillance
One of SmartPAL’s opportunities is to help instructors notice problems before grades reveal them. That should be an invitation to provide support—not an automated system for labeling students.
We aim to:
- Present signals as indicators that require human interpretation
- Avoid stigmatizing labels when a more precise description is possible
- Prefer aggregate, actionable patterns over unnecessary access to personal content
- Give instructors context and recommended next steps, not unsupported conclusions
- Measure whether interventions help students, not merely whether platform activity increases
The goal is earlier care and better teaching—not more monitoring.
5. Equity should be designed into the delivery model
AI will widen educational inequality if high-quality support is available only to students who can pay for it or to institutions with abundant staff and resources.
SmartPAL pursues:
- Universal access within participating course sections
- No student paywall for core learning support
- No public labeling of students who receive additional support
- Accessible interfaces and text-based interactions
- Support for varied devices, connectivity, and learning needs
- Language access as product and model capabilities mature
- Evaluation of outcomes across relevant student groups using aggregate, privacy-preserving analysis
Equity is not only a distribution goal. It is also a quality standard: students with the least access to office hours, tutoring, reliable technology, or informal academic networks should receive meaningful benefit.
6. Privacy and institutional governance are product features
Trust cannot be added after deployment. Privacy, security, transparency, and institutional control are core parts of the product.
That includes:
- Collecting and retaining only the data necessary for educational purposes
- Not selling student data or using it for advertising
- Not using identifiable student content to train general-purpose third-party models
- Enforcing institution, course, role, and user boundaries
- Protecting private reflections and other sensitive content according to clearly stated rules
- Keeping unpublished course materials and unreleased grades out of unauthorized AI responses
- Providing clear processes for access, deletion, and incident response
- Maintaining auditability for sensitive actions and recommendations
Institutions should understand what SmartPAL sees, what it stores, what it shares, and how long information is retained. Students and instructors should not have to infer these rules from system behavior. For current practices, see our Privacy Policy.
7. AI literacy is part of education
Students should learn not only how to use AI, but how to decide when its use is appropriate and how to judge its output.
SmartPAL should help students practice:
- Asking precise questions
- Checking sources and identifying uncertainty
- Detecting errors, bias, and unsupported claims
- Revising AI-generated suggestions
- Knowing when independent work is educationally important
- Disclosing or documenting AI assistance when course policy requires it
The student—not the AI—must remain accountable for submitted work and for judging whether a solution meets the assignment requirements.
8. Evidence should guide expansion
SmartPAL is optimistic about AI’s potential without treating every improvement in usage as proof of educational value.
Evaluation should examine:
- Learning and course-performance outcomes
- On-time completion and persistence
- Quality of student reflection and self-regulated learning
- Instructor time saved and time redirected toward students
- Whether support reaches students outside normal office hours
- Differences in impact across student groups
- False positives and missed signals in early support
- Student and instructor trust
- Privacy, accessibility, and academic-integrity outcomes
Research use should be clearly distinguished from ordinary instructional use, with appropriate institutional review and voluntary participation where required.
What SmartPAL will not do
SmartPAL will not:
- Present itself as a replacement for instructors, advisors, tutors, or human care
- Complete assignments or encourage students to bypass learning
- Make high-stakes academic decisions without accountable human review
- Treat engagement data as a complete picture of a student’s learning or life
- Expose private student content more broadly than necessary
- Substitute a generic chatbot experience when course context is required
- Require students to purchase access individually in order to receive core support in a participating course
- Optimize for automation when a human relationship is educationally important
- Make claims about privacy, accessibility, security, or outcomes that are not aligned with current practice and evidence
A practical decision rule
Before building or deploying a new AI capability, we ask:
Does this expand human capacity and improve learning, or does it merely replace human effort with less accountable automation?
Capabilities that improve learning, access, and timely human support should advance. Capabilities that increase dependence, obscure responsibility, weaken privacy, or automate away essential human relationships should be redesigned, bounded, or rejected.
In one sentence
SmartPAL uses course-grounded AI to extend the reach of human teaching and student support—helping every learner get timely, trustworthy assistance while keeping learning, judgment, privacy, and human relationships at the center.