AI Needs Context. Educators Must Decide How Much.

Key Takeaways:

Building trusted AI requires collaboration. Educational institutions, technology providers, and policymakers must work together to create AI that is effective, interoperable, and protects learner privacy.
Help shape the future of AI in education. 1EdTech's Request for Comment on trusted portable learning context invites the education community to help define how AI receives the learning context it needs, while ensuring student data remains protected.

Artificial intelligence is moving quickly into classrooms, advising offices, student support services, and administrative systems. As these tools become more capable, they can offer increasingly personalized help to learners. But that support depends on context.

Giving an AI system access to data it does not need can create unnecessary risks to learner privacy. The challenge is therefore not simply to share information or withhold it. The challenge is to provide all the information needed for a specific learning moment, but no more.

That is the challenge the principles of trusted portable learning context are intended to address, and one that educators and edtech providers must solve together. Below, we’ll explain what trusted portable learning context is and the challenges the 1EdTech community is working to address.

Think of trusted portable learning context as a learner record from a trusted administrator

The idea becomes clearer when we compare AI-supported services with the people who already support students.

A counselor may need a broad view of a learner’s academic progress, financial circumstances, and personal needs to determine what support would be helpful. But the tutor assigned to that student may only need to know which math concepts require additional work.

Each person receives the package of information necessary to perform their specific role without any additional information they shouldn’t have, from a person or system that knows the rules and processes around that data.

Educational institutions and administrators are responsible for making those decisions. They determine who may access information, for what purpose, and under what conditions, and then make sure that information is provided in the required timeframe. They must also make sure those decisions follow institutional policies, legal requirements, and commitments to students and families.

A trusted portable learning context-enabled system would help institutions apply those same responsibilities when AI agents request learner information.

Humans would continue to establish the policies. Technology would help apply those policies quickly and consistently when information moves between systems.

Why existing approaches are not enough

Most education technology integrations were built for known systems exchanging set amounts of data on predictable schedules. AI works differently. Several agents may work together in seconds, retrieving information, reviewing a learner’s progress, and creating personalized support across multiple platforms.

For that process to work responsibly, every agent must understand the same learner, purpose, permissions, and limits. Without a trusted way to provide that context, each system may make its own decisions about what data it needs, how it may use it, and how long it may keep it.

It’s not enough for an AI tool to say “we follow the necessary laws.” The institution responsible for that data needs to know where it's going and how it's being used.

A trusted portable learning context would provide a governed, time-sensitive package of information for a specific educational purpose, helping responsible parties answer three essential questions:

Does it know enough about this student to help them well?
Is it prevented from seeing what it has no business seeing?
Can anyone tell, afterward, that both of those things were true?

Six proposed principles

1EdTech is proposing six candidate principles for community review that would inform a trusted portable learning context. Together, they reflect two equally important responsibilities: protecting learner information and providing enough context for AI to be effective.

Consent

Consent helps determine what information may be released for the purpose at hand.

Sovereignty

Sovereignty means the organization responsible for the learner and the information remains responsible for deciding what is released.

Provenance

Provenance provides a record of where the information came from, who assembled it, which policy was applied, and when it was created. Adding a layer of trust.

Real-time assembly

Learner needs can change quickly. A student may complete an assignment, receive a new accommodation, change courses, or demonstrate mastery of a concept.

Querying

Querying allows an AI-supported service to request the information needed for a particular task rather than receiving a large, fixed set of learner data.

Composition

Learner information is often spread across multiple systems. Course activity may be in one platform, assessment results in another, and support information somewhere else.

Why Educational Institutions and Edtech Providers Must Work Together

These principles are not final. 1EdTech is releasing them through a Request for Comment so educational institutions, technology providers, policymakers, educators, learners, and others can help refine the approach.

Important questions remain: Are these the right principles? How should institutions define what data may be shared, with which AI tools, and for what purposes? What happens when a tool requests more information than an institution allows? Is there anything else we haven’t thought of?

The community must also determine how these principles can be put into practice in ways that are clear, consistent, and workable.

1EdTech is asking all members to review the Trusted Portable Learning Context Request for Comments document and join the discussion here to help us answer those questions and determine the best course of action to meet those needs together. Anyone interested in hosting a larger discussion or convening around the trusted portable learning context, please contact support@1edtech.org.